Some Aspects of Older Writing Systems:
With Focus on the DSS

(c) 1999, Rochelle I. S. Altman

Scans will open in a new browser window.

exodus.jpg (48K)

serech.jpg (107K)

yadi.gif (69K)


The following three-part article on the Phoenician/Hebraic Writing Systems is the result of discussions that began with the often ignored distinctions between a scribe and a calligrapher. The exchange then wandered into the importance of isolating scribal hands when examining a document; flowed from there into the alphabetic status of the Ugaritic, Phoenician, and Hebraic symbol set system, and strayed into the use and purpose of the variant forms found in the Qumran documents. The variant forms brought the seemingly odd spatializations and "elongated" graphs into focus, which, in turn, led to the trilinear limit systems used in these older writing systems.

Writing systems are systems in the precise dictionary meaning of the word: A set or assemblage of things connected, associated, or interdependent, so as to form a complex unity. A writing system comprises a series of necessary components, sub-systems, each designed to function as a self-contained unit within the overall system. These sub-systems are: a finite symbol-set system, a prescribed script system, a limits system, a mensural system, a size system, a punctuation system, a comprehension system, an orthographic system, a format system, and a content system. Each sub-system must be designed to interact with every other sub-system to create this "complex unity."

The interconnectedness of a writing system means that when we examine only a script system or a spelling system or a content system, we are creating boxes, separating the parts from the whole. Although it is much easier to examine small pieces, we must remember to put the pieces back into their appropriate places or we lose three quarters of the information. All parts of a writing system are constrained by political or religious affiliation; they were in antiquity, they still are today. Within any given writing system, be it Eastern or Western, Ancient or Modern, each component holds meaning beyond that of the written word.

While the necessary components of a writing system have not changed across the millennia, purposes can change. Modern writing systems are semantic-based; their purpose is to convey data. In antiquity, "the voice of authority" was not a metaphor to be tossed about by literary critics, but a concrete, visible reality. For more than four of the five thousand year history of writing, writing systems were phonetic-based; their purpose was to record the voice of absent authority -- be it poet or priest, judge or king.[1]

The documents from Qumran use one symbol-set system,[2] but two different script systems: Paleo-Hebraic and Square Aramaic. Both script systems are written using the trilinear limit system, the structural foundation of the Phoenician/Hebraic writing systems. In the first article, the stele of Kilamu King of Yadi is used to describe the function of trilinear limit systems, introduce the Phoenician/Hebraic writing systems, and explain the alphabetic purpose of variant forms. The second article examines this same Phoenician/Hebraic alphabetic writing system as it is used in two fragments of Exodus found at Murabba'at. Both the Yadi stele and the Exodus fragments are executed in formal authoritative scripts systems; the third article discusses the Phoenician/Hebraic writing system and variant forms used in a secular document written in an informal, cursive Square Aramaic script. To place these articles in context, it seems reasonable to summarize the discussion on scribes that precipitated them.

The series began with the very basic point that a scribe is not a calligrapher. Although both the calligrapher and the scribe use the same basic tools of the trade, the training and requirements are very different and demand entirely different mental attitudes. Calligraphers are artists; scribes are rote learners.[3] Calligraphers call attention to the writing; scribes call attention to the writer. Calligraphers are free agents; they emend texts, and elongate graphs and employ variant and/or archaic forms for artistic ends. Scribes are bound agents, in a position of fides, trust;4[] they do not emend texts, and neither extend a graph nor employ variant and/or archaic forms unless they must use them at that particular place.

One thing consistently forgotten when we discuss scribes is that we are referring neither to a little mole of a clerk out of Dickens earning a shilling a week, nor a "poor little monk" working away in some "cold and ill lit" scriptorium.[5] When we refer to scribes, we are talking about a cadre of highly trained literate people at a time when only the elite and the urbanites were literate. Further, scribes very frequently were in a position of great power and had to be reliable and trustworthy. The scribe to the Governor of Thebes, for example, received petitions -- if someone wanted to see the Governor or bring something to his attention, they had to write to the scribe.[6]

'Scribe' is actually a generic term for a very large number of occupations, many with individual equivalents today, such as: copy typist, type-setter,[7] layout and paste up professional, notary public, tax assessor, tax collector, private secretary, administrative assistant, public accountant, court recorder, and solicitor.[8] All of these occupations require fides, trust; none of them permit editorial intervention on the part of the 'scribe'. The more important the job, the less supervision; nevertheless, there is always somebody around to check on the scribe. While modern Notary Publics, for example, do not need to memorize the formulae, as ancient scribes did, they must keep records of transactions; we have just such records from Egypt on "notarial" transactions.

Whether employed in a communal capacity or working for one person only, a scribe had to be trustworthy. The penalties specifically mentioned for violating this trust were severe: Claudius, for example, mandated the death penalty for any scribe who changed the official voice in a written document by as little as an abbreviation. Another penalty was to remove the fingers of the writing hand of an offending scribe.[9] A scribe who violated his or her trust was a traitor in the fullest meaning of the word.[10] He or she could start a war, create a depression, cause a famine, put someone into bankruptcy, or send people into captivity.

The ancients were perfectly aware of human cupidity and greed. The idea that writing is to communicate comes from the Romantic movement of the 18th and 19th centuries. Most un-Romantically, writing came about through trade: the primary purpose of writing was -- and still is -- to control and record the word. The oldest extant examples of writing are to control and prevent fraud.[11]

Scribes controlled the word. Scribes as a class had to be trustworthy; but humans are humans and various techniques were employed to keep an eye on their work. On the whole, however, we are better off accepting the fact that an "editing" scribe would be emending the auctorial voice. He or she would be working in direct violation of the trust placed in him or her; the punishments meted out for such acts reflect the seriousness of the crime.[12]

When we look at ancient documents, we are looking at works inscribed or handwritten by a group of highly trained literate specialists; people who were also in a position of trust. A scribe would have to have a compelling reason to risk life, position, and security by changing the auctorial words. Certainly scribes make errors; in many cases such errors are corrected by another scribe acting as proofreader. Many mandatory forms, however, are frequently described as 'scribal errors' or indicative of "poor quality" in execution, when they definitely are not.[13] When a scribe uses a xenographic exchange, such as the insertion of Paleo-Hebraic into a document written in Square Aramaic, he or she must do so. If we see variant forms of a graphic symbol, it is because the scribe was required to use that form in that specific place. If a graph, such as a 'yod' is extended so far as to appear to be a 'vav', the scribe wrote it this way because there was no choice.

Content is important, as content determines the script, size, and format of a document. Content, in turn, is restricted by political and religious affiliations. Content, thus, establishes the context, the setting of a document, but thereafter is of peripheral importance when examining the writing system itself.

System design and system performance are two different domains. While the design of a writing system lies in the domain of system design, the execution of a writing system lies in the domain of system performance. The documents we study are examples of system performance, domain 2, as executed by scribes. Scribes are neither designers nor authors; they do not emend texts. They neither use xenographic exchange nor variant forms at random: there is always a reason. These variant forms, these extensions of graphs, these varying spatializations between words, these 'foreign-graphs', have meaning, and were placed there, not by artistic whim, but by deliberate scribal purpose. In the following three articles, we are not examining the content; we are examining the Phoenician/Hebraic writing system as performed by scribes.


Limit Systems and the Stele of Kilamu King of Yadi

Limit systems are the framework, the structural skeleton, of a writing system.[14] They are designed to distinguish semantic units and follow naturally from the decision to concatenate graphic symbols to create clusters of expressions. No matter the direction of writing (horizontally, vertically, left-to-right, right-to-left), all writing systems assign limits. The outer framework of Western writing systems consists of two horizontal lines: an upper and a lower limit. This structural foundation confines the writing zone (the area where the graphic symbols are placed) to two possible formats:

1) A central area with vertical and horizontal movement
within the zone, or
2) an entirely filled zone.

Bilinear limit systems fill the entire space between the upper and lower limits (Fig. 1a). (Our modern capitals use bilinear limits.) As the symbols fill the entire space, bilinear limits are necessarily static. The main purpose of a bilinear limit system is to confine and constrain the written word. Such limit systems intentionally "freeze" the words into an unchanging form to preserve the magical power of the word and to control people and things. (Naming is a means of control.) Bilinear limit systems are preferred by magical-mystical oriented societies. The Egyptian writing system was totally bilinear. Formal documents in Etruscan, Official Roman Imperial, and Official Neo-Babylonian also used bilinear limits.

Trilinear limits are dynamic; the words are written within the two limits. Such systems have a central writing zone with three positions: outer upper limit, upper limit of the central zone, and outer lower limit (Fig. 1b). This permits the symbols to move up and down and from side to side in imitation of the rhythms of speech. Speech itself is dynamic; duration, stress (volume), and phone (quality) change as words are vocalized. The main purpose of trilinear systems is to record words as-spoken.[15]

Upper __________________   ___________________

                                              ----------------------------- }Central

Lower __________________  ___________________ }Zone

Fig. 1: (a) Bilinear Limits (b) Trilinear Limits

In trilinear systems, the symbols hang from the two upper limits. The amount of vertical movement is constrained by the proportion of the central zone to the upper and lower limits. These limit systems allow symbols to move around within the limits. The North Semitic (both North-West and North-East) languages were, and are, stress languages, that is, stress follows the natural rise and fall of the parsing rhythms of these languages. A minimum of three levels of stress notation are necessary to record such speech: average (or normal) stress, medial (or secondary) stress, and primary stress. In trilinear limit systems, the symbols move up and down according to the stress rhythms of the languages.

Durational notation, that is, the length of time a sound should be held, is recorded by the amount of movement from side-to-side, that is, expansions and contractions of the space between graphic forms.[16] This accounts for the seemingly odd clumping and spacing of words in texts written or carved in trilinear limit systems.[17] Ligatured symbols, for example, indicate phones that are spoken in close succession, while spaces, which are measured in terms relative to the width of an 'ayin' (the durational mensural basis in Semitic systems; an 'o', in Latin and Greek systems[18]) indicate the length of time a sound should be held when read or sung or chanted.

The ancients knew that sound is relative and varies depending upon whether one is, for example, a baritone, an alto, or a soprano. They also knew that the auctorial voice is a constant. Hence, sound is a variable; voice is a constant. They knew that sound (the variable) could not be written down, but that 'voice' (the constant) could be.

There is one very important point that we must not forget when looking at ancient documents: Silent reading is a very late and very modern concept. Silent reading conditions one to the concept that the only way to render the auctorial voice is statically. In antiquity, the written word was meant to be performed, executed out loud -- as recorded, whether bilinear or trilinear limits are employed. The two limit systems -- the one static and the other dynamic -- are quite different in intent. Bilinear limits restrict and confine the spoken word; trilinear limits permit and support the rhythmic ebb and flow of speech as spoken. The primary purpose of both limit systems nevertheless is to convey how to say the written record.

In the West, we are conditioned to the Roman inscription system, that is, totally bilinear, static, confined, and with no allowance permitted for relative variables. As inscriptions from late 6th century BCE, such as those commissioned by Darius I (Fig.2), are also totally bilinear, when confronted by inscribed documents, such as the Yadi or Meshe stelae, the assumption has been that these 9th century inscriptions are "primitive" and that the carvers were still "learning."

------------------------------------------------------------- II III XI I<- -IX <II \ <<II <<II III -------------------------------------------------------------
Fig. 2 The bilinear limit system on the Darius inscription from the late 6th BCE

Stelae inscribed in the Phoenician system are hardly primitive: even if the Yadi stele had not been commissioned by a king, we merely need look at the very carefully carved round point word dividers to know this. King Kilamu of Yadi's inscription carver was not "learning"; he was using a different writing limits system. Unlike the mystical-magical "frozen" bilinear limit system, the North-West Semitic pragmatic-practical writing systems use the "flexible" trilinear limit system. In addition, just as the Ugaritic Alphabetic cuneiform symbol-set uses variant forms to record separate phones, the Phoenician alphabetic symbol set also uses variant forms to record separate phones.

We shall now look at the scanned image of the stele of Kilamu King of Yadi (yadi.gif).[19] Please examine the examples in the box on the left-hand side of the scan first; they are for orientation purposes.

Example 1 isolates the three 'aleph' phone graphs used in the Phoenician alphabetic symbol-set to record the different vowel phones attached to a given graph.[20] Note that the graphs all have the same basis, but are actually quite different. The first symbol, the tilted aleph with a "standard" minim (the leg) running from the outer limit to the upper edge of the central zone, is the "standard" 'aleph'. The middle graph, (which is dropped down for stress notation) has an upright minim and the legs of the "lobe" (the round part of an 'a', referred to as the "bow" in a 'p' in "Latin" graphs) are close together. The third 'aleph' is short, somewhat rounded (quite rounded in cursive) and, unlike the other two forms, the "lobe" is almost nonexistent while the legs are clearly delineated.

Example 2 shows two "standard" alephs (line 9) with a point word divider between them. The leg on the right-hand 'aleph' has been extended all the way down to the lower limit -- indicating primary stress.

Example 3 shows an unstressed standard aleph and a stressed 2nd form (line 7).

Example 4, from line 10, directly under Kilamu's feet, gives a clear example of stress notation in action.

If we now look at the stele commissioned by Kilamu King of Yadi, we can see the Phoenician trilinear limit system in action. The majority of the symbols hang from the outer upper limit -- which is to be expected: normal or average stress predominates in most stress languages. Stress languages are quite intolerant of unstressed speech. When somebody speaks with too many stressed or unstressed syllables in a row we say they are "mumbling." Look carefully, and we can see that, even in line 1 where the outer limit is not inscribed, the carved symbols vary in the amount of "drop" from an imaginary outer limit. This type of "drop," movement up and down from the outer upper limit, can be seen on every line of the stele.

We can also see that the amount of kerning (the space between symbols) varies as well. Again, this is to be expected, as someone "orating" does not speak in evenly spaced utterances, such speech is what we call a "monotone" and is guaranteed to put an audience to sleep -- whether in a lecture hall or when telling a story. Durational variation shows in the clumping and spacing of the letter symbols and occurs on every line.

The closest modern analogy to the Phoenician/Hebraic writing system is a visual tape recording -- including spectrophonography[21] -- these writing systems contain everything on how to reproduce, relative to the variable (the sound), the auctorial voice (the constant).

The above may be verified for him or herself by anybody who wishes simply by following the guidelines given and by examining the stele of Kilamu King of Yadi. The writing system used on the Yadi stele is the same one used in the Paleo-Hebraic texts from Qumran. The trilinear limit system, durational and stress notation, as well as numerous variant forms are also used in the texts written in Square Aramaic.


Two fragments of Exodus from Murabba'at

In the Yadi stele, we examined the trilinear structural basis and looked at one of the variant forms used in the Phoenician-Hebraic writing systems. We can see that variant forms are just that, variations on a basic graph; we also saw that these variations can be quite subtle -- as slight as a straight-back or closely spaced 'legs'.[22] The Yadi stele shows us how trilinear limit systems work: durational notation is indicated by movement (clumping and spacing) on the horizontal plane; stress notation is indicated by movement (up and down) on the vertical plane.

The textbooks tell us that stress is grammatically determined in Hebrew, and this is true. What the textbooks do not tell us is that what is true in theory is not always true in practice. Stress is very flexible and a specific syllable may receive none, secondary (medial), or primary stress - even in Hebrew. Stress can either be intrinsic, that is, the normal (textbook) pronunciation of the word, or extrinsic, that is, impressed by musical, poetic, or rhetorical requirements -- of which emphasis is the most common. In other words, any document that records quoted statements may or may not follow the intrinsic (normal) rules for pronunciation. This is an important point to remember when examining these two fragments from Murabba'at; both Exodus fragments are quoting statements: the first spoken by Moses to Aaron and Aaron to the elders, and the second spoken by God to Moses. Fragment number 4 (above on the scan) contains part of Ex 4:28-31 and fragment number 5 (below on the scan) contains the last syllable of Ex 6:5 and most of Ex 6:6. (DJD II, Plate 1 - for those with access to the DJD.) What is written on the fragments is what is termed 'content'.

As has already been noted, content determines the correct script, size, and format of a document - whether inscribed or written. Content, in turn, is constrained by political and religious affiliation. For these reasons, once the content has been examined to create a context -- cultural, political, or religious (or any combination of the three) -- content is thence forth relegated to a part of the writing system as a whole.

While formulae that may be used to reconstruct the original size of a damaged document are recorded in scribal handbooks[23], we need at least one full margin (any of top, bottom, or inner) plus one full sentence to apply them. These fragments are too small; while we can see that the document is written in a narrow column format, we do not have even one margin -- size cannot be retrieved. The script, however, is there for us to see.

These fragments are written in a very formal script. Content tells us that this formal script is an authoritative/official biblical font, as do the large number of graphs requiring multiple strokes (pen lifts). Further, although written with a square-cut pen nib and an essentially monoline design, there are many wedges (the 'chip' of a starter stroke made by starting just below and to the side of the beginning of the main stroke), imitative of the cuneiform thick-wedge to thin-leg starter strokes, as in the central stroke of the 'sin' in Fragment 5 or the starter stroke of the 'lamed' on both fragments. (Serifs are mutations of these wedges.[24]) Each wedge is an additional stroke. Formal authoritative and/or official script designs incorporate many pen lifts; these pen lifts have one primary purpose -- to slow the scribes down in order to protect the authoritative words from being garbled in transmission.

There is one other point that should be mentioned about this type of formal script design. Ductus (the direction of a pen or brush stroke) cannot be used to isolate scribal hands in these scripts and fonts. Unlike cursive designs, where ductus may be determined for a given scriptorium (but not necessarily for one scribe), the ductus on formal scripts and fonts is constrained by the design of the script. The only way to isolate hands on documents written in an official script or font is by scribal ideographs.

Only senior scribes have ideographs. A scribe learns a font by practice until his stored patterned nervous response is totally automatic. Once a font has become part of the patterned processes in the motor area of the brain, a scribe acquires ideographs. Scribal ideographs do not change; the scribe literally cannot help but write these forms this way. His responses are so automatic that he associates a font with the feel of a pen. Until the response is completely automatic, scribes do not, cannot, have ideographs. These virtually robotic responses allow us to isolate scribal hands.

Anyone who has worked with any pre-printing press or hand-written documents knows that scribes are not machines. Even in something as carefully executed as The Book of Kells[25] there are differences, and the cursive square Aramaic used in so many of the Qumran texts will show much more individualism. What does not vary, in terms of isolating hands, is the angle of attack, an individual's tendency to leave graphs open, or a manner of writing "finishing" strokes.

These two fragments of Exodus display scribal ideographs; each fragment was written by a different scribe. When we look at the scanned images ( we can compare, for example, the 'bets', 'nuns', 'yods', and 'sins/shins' on Fragment 4 (Ex 4:28-31) with those on Fragment 5 (Ex 6:5-6).

This formal script design requires a four-stroke 'bet' (starter wedge, head [cross-stroke], minim [the upright "leg"], and an angled bottom stroke that crosses the right-hand leg to create a 'serif'), a three-stroke 'nun' (starter wedge, minim, and bottom stroke ending in a 'serif'), and a two-stroke 'yod' (starter wedge and minim).

The word 'BNI' appears on both fragments; 'BNI' is a bound form. Bound forms are a major source of scribal ideographs: the scribes literally cannot help but write the forms this way. While both scribes use stress notation (the elongation of the yod is stress notation in action), there are clearly visible differences. The scribe of fragment 5 starts with the merest 'chip' of a wedge approach stroke and writes his head stroke almost straight before coming down to make the minim; all three strokes are written in a single continuous operation (which has led to this being erroneously called a "two-stroke bet."). He does not write an obvious 'serif' on the bottom stroke of his 'bet', which, as the head stroke, is almost straight. While technically his 'bet' can be called a four-stroke bet, he does not use the extra pen-lifts that are part of this design.

The scribe of Fragment 4 uses a much sharper angle of attack in his approach wedge stroke and a much steeper angle on the head stroke of the 'bet'. While he writes the three strokes (starter wedge, head, minim) in an almost continuous movement, he nonetheless maintains the appearance of the four strokes, including finishing his bottom stroke with the correct-for-this-design serif. The 'bet-resh' in line 4 of Fragment 4 shows exactly the same 'bet' ideograph, in other words, this is the way this scribe writes 'bet'.

The 'nuns' are also different. The scribe of Fragment 4 writes a steeply angled starter wedge, three-stroke 'nun' with a steeply angled, straight line lower stroke in opposition to the scribe of fragment 5 whose rounded, two-stroke 'nun' (starter wedge and minim/bottom stroke combined) does not belong to this squared, formal script design. On fragment 5, the 'nun-yod' of 'ANI' is not ligatured; but the 'nun' swings down below the lower limit. (This is stress notation again.)

Both scribes ligature all three graphs. The scribe of fragment 4 makes the wedge on the right-hand side of the bottom stroke of the 'nun' touch the left-hand of the bottom stroke of the 'bet'. His stressed 'yod' is not only curved, instead of straight-backed, but comes all the way down to touch the left-side extension on the 'nun'. The rounded bottom corner stroke of the 'nun' touches the unserifed bottom-stroke of the 'bet' in the hand of the scribe of fragment 5. As the scribe of Fragment 4, the scribe of 5 makes the slightly rounded minim of his 'yod' touch the bottom stroke on the left-hand side of the 'nun'.

In other words, the differences between the 'nuns', 'bets', and 'yods' on these two fragments are scribal ideographs; they are neither the vagaries of one scribe nor are they variant forms. Variant forms, as we saw on the Yadi stele, are used to indicate different phones. Variant forms do occur on the 'sin/shin'. The font design calls for a six-stroke 'shin/sin', with individual starter wedges on each of the three minim strokes. Both scribes have evolved techniques to "speed up" the formation of the graph.

On Fragment 5 there is only the one 'sin' in the word 'Israel'. This scribe uses a combined wedge starter/right-hand leg stroke that, by rounding the leg to the extent that it approximates the right-hand side of a lobe, turns the two strokes into one. The central stroke of his 'sin' is thick at the starter wedge and ends in a thin "point" that meets the left-hand leg (without starter wedge) above the base. Thus, the Scribe of Fragment 5 produces a three-stroke 'sin', which, nevertheless, retains the formal, squared, aspects of this script design. The 'shin/sin' of the Scribe of Fragment 4 exhibits very different attributes.

As we might expect from the techniques employed to retain the appearance of writing a four-stroke 'bet', the Scribe of Fragment 4 uses a wedge starter stroke on each leg of his 'sin/shin' graphs. He then writes each wedge and minim in a continuous stroke. His starter wedge stroke is again at a steeper angle of attack and the right-hand minim is much straighter than that of the Scribe of Fragment 5. He also dispenses with the more time-consuming variant form differentiation of wedge-starter, point-above-base that we see in the hand of the Scribe of Fragment 5.

On fragment 4 we have a number of 'shins' in line 2 (moshe, shel-, asher). On this line, the scribe always keeps the central stroke of the 'shin' within the outer legs, but there is no question that in 'shel' and 'moshe' the central stroke comes right down the center, while the 'shin' in 'asher' meets the left-hand leg above the base. On the other hand, in line 3, the 'sin' in 'Israel' extends well below the base. On line 4, the 'shin' in 'asher' appears to also extend below the line, but close examination shows that this is not a purposeful extension, but a blob of ink undoubtedly caused by the scribe's short-cut technique and the speed with which the form was executed. On line 5, the 'sin' in VIAS again extends well below the line.

There are clear differences among graphs in both fragments, particularly on the aleph, vav, and shin/sin, which strongly suggests variant phones. We can see that, for example, there is a definite difference in the length of the trailer on the left-hand leg of the aleph (e.g. segol-aleph in 'emor' and 'et' vs. tzere-aleph in 'Israel' on Fragment 5; segol-aleph in 'et' vs. tzere-aleph in 'Israel' on Fragment 4). In both fragments, the extension on the right-hand leg of the 'aleph' in 'Israel' is stress notation, not variant form notation.

The fragments are too small, however, for us to map graph to phone.[26] Both scribes, nevertheless, use the Phoenician-Hebraic trilinear limit system that we saw demonstrated on the Yadi stele, and both use stress and durational notation.

There is one other point, that may seem obvious, but should be stated nonetheless. While the oral tradition behind these texts has been studied by many, the implications of orality lying in the background, who knows how far in the distant past, have not been faced.[27] These words are not meant to be read silently; they are meant to be chanted. We cannot overemphasize the importance of music and chants to the act of reading out loud. Chants are a very effective mnemonic aid; a fact that was known back at Kesh and used by Enheduanna of Akkad in her religious music set.[28] Music/chant is an integral part of the transmission of these documents -- whether secular or sacred.[29] The fact that among the various branches of Judaism the only differences in the chant assigned to the Pentateuch are in the musical tradition and not in the melodic motif, tells us that the melodic motif was already firmly established centuries well before the turn of the Common Era. When reading any of these documents, we should read the text out loud as if we were reading a musical score: one 1/8 note per syllable.

As we go through the fragments marking the notations, it may be amusing to compare the reality of the words as chanted[30] with the printed "textbook" record in the BHS.

Note the clumping and spacing of durational notation - graphs run together ligatured, or space inserted between graphs that are normally part of a word. Primary stress graphs are noted by a carat (^) under the graph, secondary stress by a plus (+), ligature by an underline (_). As too few people know Michigan-Claremont to use in a general posting, 'X' = chet, 's' = samech, Sh = shin, 'n' = nun-sofi, Tz = tzadi, 'a' = ayin.

Fragment 4: Ex 4:28-31 The first graph on this fragment is the nun-sofi in the name 'Aaron'.

Line 1:   n A_TKL DB_RI IHVI A (part of the aleph in AShR, the next word)
    (BHS:   AT KL-DBRI IHVI)  

Line 2:   ShLXV VA_T KL HA_T_T AShR Tz_VHV VILX MSh H (part of heh missing)
                  +   +                     +

Line 3:   VAHRn VIAs_P V AT_KL ZKNIBNI ISRAL VIDBR (part of resh missing)
                 ^    *  +        +  ^    ^   +
* Note the primary stress on the 'yod' and the extra space between the 'pe' and the 'vav' indicating a "held" sound on VIAs_P V. We should bear in mind that these words were chanted, not read silently. It may be clearer if we think of each syllable as the equivalent of an 1/8 note when sung out loud.
Line 4:   AHRn A_T KL HDB_RIM AShR DB_R YHVHAL M (part of mem missing)
               ^            ^               +

Line 5:  [VIa]S HATT La_INI Ha_M VI[MAn]      [missing parts] 
                        ^ +   ^   + 
    (BHS  VIaS HATT LaINI HaM: (31)VIMAn  

Line 6:  [] IHVH A_T BNII[SRAL]              [missing]

Fragment 5: Ex 6:5 and most of Ex 6:6  

Line 1:  []TI LKnAMR LB_NI ISRAL ANI YHVH       [missing]
                 +       +    ^   ^^     

Line 2:  []Tz_ATI A_TKM M_T_X T  sBLT MITz_RIM  [missing]
                      +                   +     

Line 3:  []L_TI A_TKM Ma[BD]T_M VG_ALTI A_TKM [letters cut off at bottom]

Note: As the system calls for extension downwards, we cannot determine the stress notation for most of line 3.

We can see that, while the words are the same in both the BHS and the two Fragments, what is considered one hyphenated word in the BHS is quite different in the originals. It may also be noticed that the extra stress on "Ani" in line 1 makes sense: This particular stress notation accords with the same type of stress on ANI that is chanted in the Sefer Haggadah. Another point worth mentioning is that whether words are separated or run together depends upon the parsing rhythms of the speaker and whether he or she is emphasizing something for rhetorical effect.

While these older writing systems do take some time to adjust to, once we can read the words as spoken, we will have a much better understanding of the texts. In addition, it is also a great deal of fun to see the words literally 'speak' to us from the past. The pages have voices -- IF we will listen to them.


"Haserech" from Cave 1 at Qumran
(c) 1999, R.I.S. Altman

Both the Yadi stele and the Exodus fragments are written in formal, official scripts. We also saw the Phoenician-Hebraic trilinear limit system and variant phone system in use. We are now going to look at something quite different: a third class copy of a document written in an informal cursive font -- a mutation of the Square Aramaic. For convenience, the document shall be referred to as 'haserech."

Content, as we have noted, determines the script, size, and format of a document. In the Hebraic hierarchy of formats, official and authoritative documents are written in narrow columns, such as those we saw on the Exodus fragments. (We can tell the width from Fragment 4 where we have almost complete lines of text.) Unofficial, non-authoritative, or informal documents, such as commentaries, are written in broad columns. (The Alexandrian-Roman Christian parties reversed this order: broad for authoritative, narrow for unofficial.)

Ancient writing systems also have a hierarchy of sizes: the largest documents are always 'The Law'. The Paleo-Leviticus, for example, after placement of fragments as to where they should be, shows it to have been around 22" (or more) in height, including margins. 'Writings', on the other hand, run around 10-1/4" to 11" in height. Q11Ps, the large Psalm scroll, is around 10-1/4" (26.2 cm) high, as is the large Isaiah scroll. (Christian hierarchies varied depending upon affiliation; the largest size among all early groups is always 'The Law' (The Pentateuch); later, the Gospels and then complete Bibles appear as the same size; secular law codes are second in height with 'writings' (K'tuvim'] relegated to third place, and so on down the line to the 'tablet' size which means "entertainment.")

Finally, these systems had an ordered hierarchy of scripts both formal and informal. There are formal "biblical" fonts for 'The Law' (e.g. Paleo-Hebraic and the very square, wedge, Square Aramaic of the Exodus Fragments); several formal 'liturgical' fonts for 'writings' differentiating between class of 'writings' (e.g. the "simple" Square Aramaic of Isaiah A vs. the highly stylized font used in Q11Ps); formal official 'secular' fonts (the so-called 'Herodian' is such a font), and official 'chancery' type fonts - such as that seen on the Papyrus letter from the administration of Bet Mashko and found at Murraba'at. The majority of the non-Biblical documents are written in informal cursive fonts that vary, depending upon the level of book class (and time frame), from the 1st class copy of the "Children of Light and Children of Darkness" work, to the careless third class copy of 'Haserech'.

The first item of importance that we can see when we look at the photograph of 'Haserech' is that this document is written in the broad column format as opposed to the narrow column format we saw in the Exodus fragments. This already tells us that this is not an official or authoritative text. The next things we notice is that the document displays all the signs of a hurry-up job.

The word 'incipit' means "it begins" and refers to the opening words (or music) of a composition.[31] In antiquity, books were catalogued by the first meaningful words, a practice started by Enheduanna back at Akkad.[32] (The Pentateuch is organized by 'incipit'.) The incipit on 'Haserech' is written in the same font and only slightly larger than the body text: a sure sign of a third class copy. The incipit, nevertheless, tells us that the book was known during the period by the title 'haserech lekol eidat israel'.

The script itself is a mutation (scripts do NOT develop, they mutate)[33] of a formal cursive. We can also see one standard scriptorium practice in use: when a job requires more than one scribe, all scribes assigned to the work take turns writing the first lines. This practice is designed to obscure distinctions among scribal hands. While this hide-the-difference-technique is effective, scribal ideographs, nevertheless, defeat the technique -- if, that is, we know enough to look for the ideographs.

The scribes of "Haserech" all have distinctive ideographs; there are four hands on this column. We will look at two ideographs for each scribe (there are many more).

Scribe 1 writes his sin (l.1 Israel) with a straight central stroke and his shin (l.2 mishpat) with a curved central stroke. Both the right- and left-hand legs of his shin/sin are straight (not bowed). His 'taf' (l.1 zot) has a very heavy upward tilted headstroke caused by overwriting his first stroke: he runs back over the attack (starter) stroke and then comes down to make the right-hand leg; he then makes another starter stroke at the left hand edge swings back and then comes straight down. This leaves a sideways tilted stroke on the upper left-hand side of the graph.

Scribe 2 writes both his 'sin' and his 'shin' with a straight central stroke; however, he writes the right-hand leg of his 'shin' straight and his 'sin' bowed. His 'taf', as that of Scribe 1, is again written in two strokes, but there is quite a difference between the 'tafs' of Scribes 1 and 2. Scribe 2 starts his right-hand side with a very rounded headstroke, similar to the right half of a sans-serif 'h'; his completion stroke starts above the headstroke, comes down straight, and then makes a distinct left-hand curve out at the bottom (l.2 just before the tear - BRITM).

As Scribe 1, Scribe 3 writes his shin with a curved central stroke -- only he also curves the right-hand leg (l. 3, his third word 'sm-'). He writes his sin with an elongated and angled left-hand stroke (l. 3 on the tear). He also writes his taf in two strokes. He starts at an angle and then goes from the left-hand side of the graph, across and then straight down. He then writes his left-hand leg starting at the point where the little "flip" of his starter stroke levels out, comes down and ends in a left-hand curve.

Scribe 4's taf is very similar to that of Scribe 2; however, his left-hand leg comes straight down before ending in a slight left hand finishing stroke (l. 5 second word). Only one sin/shin appears in his first stint.

Scribe 1 wrote the incipit, all of line 1 and the first four words on line 2. Scribe 2 wrote from the fifth word on line 2 through the third word on line 3; Scribe 3 wrote from the fourth word on line 3 to the third word on line 4; and Scribe 4 wrote from the fourth word on line 4 to the first two words on line 5. If we look at the photograph, we can see that this order creates a staggered overlap of scribal hands, each of approximately the same length. Scribe 1 then wrote from the third word on line 5 through the seventh word on line 7. Now Scribe 3 took over. The entire column is written in stints that last as long as the ink holds out: approximately two to two and one-half lines per scribe. The order of the scribes varies in order to avoid creating a repetitive pattern that would defeat the purpose of the change of scribe technique.

If we wish to see interspersed shins and sins, we need only look for the 'taf' of Scribe 1, then check his shin/sin use. Scribe 1 is back again in line 10, 'shnat', right after the tear. He wrote all of line 11 and the first two words on line 12. He comes on again on line 13, 8th word (again 'shnat') and stayed on through 'Israel' on line 14. He shows up again on line 21, 8th word through line 22, seventh word.

This change-scribe-at-ink-end technique gives us quite a bit of information about the scriptorium. First, it was at least of medium size and probably had 8-10 scribes on hand. Second, this is a technique used centuries later at the more sophisticated scriptoria; we are not dealing with novices. Third, while the scribes write the same font, we can tell that there is an age range involved. Scribe 1 appears to be older and to have learned a somewhat different font: the forms used by Scribe 2 have mutated from an older font design. Scribe 3 also tells us that more than one font was used at this scriptorium: his deeply angled stroke on the sin is very similar to the formal cursive used in Q11Ps. It does not belong in this font design.

This sophisticated end of ink technique is intended to hide differences among scribal hands. As we have seen, it is still quite effective. Even more effective, however, is the ancient trilinear limit, alphabetic, Phoenician/Hebraic Comprehensive Writing System.


1 Quoted from the "Preface" to Rochelle I. Altman, Absent Voices: The Story of Writing Systems in the West. forthcoming, 2000.

2 For the clearest exposition of what is meant by symbol systems and bound and unbound forms, see, Chapter 1, Joseph R. Shoenfield, Mathematical Logic. Reading, Mass: Addison-Wesley, 1967.

3 Scribal training is purely rote learning; it requires acquired skill, not artistic capabilities. "Modern research into nerve-motor response patterns brought about by the space exploration program only verifies what the ancients already knew from empirical observation. In order to perform a physical action effortlessly, without thought, the nervous system must learn to respond automatically to that particular set of actions. Playing badminton will not help someone's tennis game; the stored, patterned nerve responses are different. Riding a bicycle will not help someone drive a truck, nor will running a drill press help someone learn to use a hand drill. Each activity has its own stored patterns and must be acquired as a separate set of nerve-pattern responses." (Absent Voices, Chapter 11) The acquisition of stored motor-response patterns is the foundation of scribal training.

4 Scribes as notaries and in a position of trust date back to the very beginning of the technology we call 'writing'. On the Semitic side of the ANE, contracts were required for transactions at a very early date. In some areas, it was law, in others merely custom. Custom, however, rather quickly takes on aspects of law stronger than the written law. Back at Sumer, they had a place for everything and everything was in its place. Archaeologists digging around there found baskets of clay tablets: each basket was neatly indexed by content -- and the contents of each basket were of the same type of transactions -- and the same size. Later, under Hammurabi, every transaction, no matter how small, had to be in writing and in detail.

5 The quoted words are from a specific work; however, there are so many references or descriptions along this line that it would be quite unfair to single out one person.

6 There are many extant petitions to scribes in their various roles. P. Vindob. 2 is the upper half, and P. Wash. Univ. 445, is the lower half of a petition dated 28 October-26 November, 21 BCE to the Royal Scribe, Herakleides. For other examples, see among others, K. A. Worp, Greek Papyri from Kellis: I, Oxford: OUP, 1995, or H. J. Wolff, Das Reht der griechischen Papyri Ägyptens in der Zeit der Ptolemäer und des Prinzipats. Munich: 1978).

7 Apprentices, beginners, learn by copying. The 'copyist' function is the most basic and elementary level of scribal employment. Whether a scribe went on to more advanced training depended upon the capabilities of the individual.

8 The British still differentiate between a 'solicitor' who writes up contracts, and a 'barrister', who stands before the bar.

9 The specific statement of a penalty is uncommon for the very simple reason that an offending scribe would have been prosecuted under the name of his or her act, for example, sedition.

10 When we come across a scribe writing deliberate forgeries, as in 11th century English charters such as those from Worcester, the scribe (in this specific case a man named Hemmings), was only doing his job for his employer: a monastery trying to grab what it could in the confusion that followed in the wake of the Norman conquest.

11 Forgery and fraud were (and are) always serious problems and many methods and techniques were devised to keep things under control -- from seals to filling tablets from edge-to-edge (so as to allow no room for changes), from baking official documents to prevent emendations to from requiring a specified amount of margin on the right and cutting off next to the text on the left on papyri receipts, from particular ways of folding letters to the clever Roman practice of requiring that all correspondence with respect to a complaint be written on the same page. (This is why so many of the papyri have a large, empty lower margin.) Jewish law is unusual in that it required a contract be in the hand of the party and is rather obviously yet another fraud prevention technique. For a discussion of the oldest extant writing, see Denise Schmandt-Besserat, How Writing came about. Austin, TX: U of Texas Press, 1996.

12 Although most scribes were employees, there certainly were independent scribes. Scribes opened "bookshops," but a literate public is necessary for a book-producer to earn a living. Such bookshops appear in urban centers such as Rome -- and later disappear in the West until the 14th century CE. An advertisement for a team of inscription-scribe plus stonecutter-scribe shows up from Pompeii, and, of course, there are always calligraphers. The independents did not have direct supervision, instead they had to please the consumer. If the books were not what the customers ordered, the bookshop went out of business; if the inscriptions did not please -- close up shop; if the calligraphed notice was not exactly what was ordered -- find another way to put food on the table.

13 It seems unfair to pick one person out to quote when so many remarks of this nature appear in every field that deals with old documents, but the following remark is typical of these comments and illustrates the point. In referring to the "poor quality" of the handwriting used in a catalogue written by Henry Kirkestede, a Precentor, a member of a group of musical specialists who continued to use the phonetic-based writing system until well into the 16th century, Richard H. Rouse maintains that the writing is "quite uneven; some of his letters are larger than others. He uses two different forms of 'A'." ("Bostonus Buriensis and the Author of the Catalogus Scriptorium Ecclesiae," Speculum, XLI, No. 3, July 1966. 471-99.

14 The only discussion of limit systems in particular, and writing systems as a whole, the author has ever seen was in a pamphlet, handed out to novice script designers entitled, "The Basics of a Writing System." The pamphlet may have been issued under the auspices of the WPA sometime during the 1930's. (Please excuse the lack of detailed bibliographic information; more than 40 years have passed since the pamphlet was "borrowed" -- and never returned.)

15 Modern Western limits are quattrolinear, which is merely trilinear limits moved downwards to accommodate ascenders (the "high" part of symbols such as b, d, l). In quattrolinear limit systems the symbols are written between the four lines.

16 Each graph has a 'frame', that is the amount of white space surrounding a graph. This white space is quite flexible, but can never stray too far beyond a certain point. If this space is either too contracted or too expanded, comprehension decreases. For further discussions on 'frames', see John DeFrancis, Visible Speech: The Diverse Oneness of Writing Systems. Honolulu: University of Hawaii Press, 1989.

17 Only five scholars have addressed the question of durational and/or stress notation: Robert D. Stevick, Andre Crépin, and Katherine O'Brien O'Keeffe in Anglo-Saxon Studies; Columba Kelly in Gregorian Semiology, and myself. While the documents are later, they are nevertheless relevant to the Qumran materials as the early Christians adopted the Hebraic writing system that we see in use at Qumran. These works are as follows: Robert D. Stevick, Suprasegmentals, Meter and the Manuscript of Beowulf. The Hague: Mouton (1968); Andre Crépin, Beowulf: Edition diplomatique et texte critique, traduction française, commentaires et vocabulaire. 2 vols. Göttingen: Kümmerle Verlag, 1991; Katherine O'Brien O'Keeffe, Visible Song: Transitional Literacy in Old English Verse. Cambridge: Cambridge UP, 1990; Rev. Columba Kelly, O.S.B. The Cursive Torculus Design in the Codex St. Gall 359 and its Rhythmical Significance: A Paleographical and Semiological Study. St. Meinrad, Indiana: Abbey, 1964. Rochelle I. Altman, Psalms from the Paris Psalter: Psalm 23(22). Harrisburg, PA: Lengyl, 1993. (Limited edition and out of print -- copyright reverted: parts included in Absent Voices, forthcoming.)

18 For discussions on mensural bases, see, Alexander S. Lawson. Anatomy of a typeface. Boston: Godine, 1990; Hermann Zapf. About Alphabets: some marginal notes on type design. Cambridge, MA: MIT Press, 1970; Donald Jackson, Histoire de l'écriture. Paris: DeNoël, 1982.

19 The scan of "The Stele of Kilamu King of Yadi" is Fig. 45, on page 55 in the English version of Joseph Naveh's Early History of the Alphabet: An Introduction to West Semitic Epigraphy and Palaeography. Jerusalem: Magnes, 1982. This illustration was chosen for two reasons: 1) it is quite clear, and 2) Naveh's book should be available in large Public libraries so that people can look at the illustration (and hopefully read the text) for themselves.

20 There are also variant forms of 'heh', 'vav', etc.

21 For an illustration of spectrophonograms, see the frontispiece to DeFrancis, Visible Speech, 1989. For a discussion of spectrography in relationship to sound as we see it in the old documents, see Giovanni De Poli, Aldo Piccialli, and Curtis Roads. Representations of Musical Signals. Cambridge, Mass.: MIT Press, 1991.

22 For a detailed discussion of the subtleties involved in script design, see, Douglas R. Hofstadter, "Metafont, Metathematics, and Metaphysics," in Metamagical Themas: Questing for the Essence of Mind and Pattern. New York: Basic Books, Inc. 1985.

23 The proportions of margin to white space to text space in quality "books" have not changed in more than 2500 years. See Jan Tschichold, The form of the book: essays on the morality of good design. Trans. by Hajo Hadeler; ed., with an introduction, by Robert Bringhurst. London: Lund Humphries, 1991. Tschichold worked out the formulae for himself; the 9th century handbook containing the pertinent information was not discovered until after his death. The handbook verifies the accuracy of Tschichold's calculations.

24 To this day, for a font to be authoritative, it must have serifs, a sans-serif font is taken as "advertising" or "entertainment. For a study of the importance of the serif in later documents, see Stanley Morison, Politics and Script: Aspects of authority and freedom in the development of Graeco-Latin script from the sixth century B.C. to the twentieth century A.D. The Lyell Lectures 1957. Edited and completed by Nicolas Barker (Oxford, 1972).

25 The design of the font used in The Book of Kells incorporates wedge-starter strokes. These wedge strokes are a remnant of Cuneiform script designs and one of the writing system markers that may be used to trace the movements of Early Christianity.

26 For a model on how to map graph-to-phone, see James W. Marchand, The Sounds and Phonemes of Wulfila's Gothic. The Hague: Mouton, 1973.

27 For a better understanding of the implications of oral/memorial cultures, see Mary J. Carruthers, The Book of Memory: A Study of Memory in Medieval Culture. Cambridge: CUP, 1990. (Reprt. 1993, 1994).

28 For information on Enheduanna and her organization of religious books and music, see Gerald H. Wilson The Making of the Hebrew Psalter. Chico, CA: Scholars Press, 1985, especially Chapter 1. The antiquity of these techniques may be observed in Biblical poetry. For basic readings on Hebrew Poetry, see Wilfred G. E. Watson, Classical Hebrew Poetry: A Guide to its Techniques. Sheffield: JSOT, 1984; and Willem Van der Meer and Johannes C. de Moor. The Structural analysis of Biblical and Canaanite Poetry. Sheffield: JSOT, 1988. 1-61.

29 The endurance of ethnic melodies must not be underestimated. We have ample evidence to date the known Torah chants to the Second temple period -- if not earlier. Among English charm chants, for example, the children's charm to chase rain away, "Rain, Rain, Go Away" is sung to the same chant as the Old English "Charm Against a Wen." (It must be admitted that the old words fit the melody better than the new ones; however, this is also true of English ballad melodies, such as the 13th century "I have a Swester" [sister] that resurfaces as our modern "Riddle Song.".) Many of the English melodies, some dating back to the 6th century, are still in use: today we call them nursery rhymes.

30 The melodic aspect of orality must be taken into consideration when examining Hebrew prose, poetry, and song.

For discussions on music as a mnemonic, see James, G. Southworth, Verses of Cadence: An Introduction to the Prosody of Chaucer and his followers. Oxford: Blackwell, 1954; and John Stevens, Words and Music in the Middle Ages: Song, Narrative, Dance and Drama, 1050-1350. Cambridge: Cambridge UP, 1986. For discussions on the origin of chants and the relationship between the Jewish and Christian chants, see Hanoch Avenary, Studies in the Hebrew, Syrian, and Greek liturgical recitative. Tel-Aviv: Israel Music Institute, 1963, also Encounters of east and west in music: selected writings. Tel-Aviv: Faculty of Visual and Performing Arts, Dept. of Musicology, Tel-Aviv University, 1979; Edith Gerson-Kiwi, Migrations and mutations of the music in East and West: selected writings. Tel-Aviv: Tel-Aviv University, Faculty of Visual and Performing Arts, Department of Musicology, 1980; Eric Werner, The sacred bridge; the interdependence of liturgy and music in synagogue and church during the first millennium. London, D. Dobson; New York, Columbia University Press, 1959-1984; James McKinnon, Music in early Christian literature. Cambridge: Cambridge University Press, 1987; and Willi Apel, Gregorian Chant. Bloomington & Indianapolis: Indiana U Press, 1990. (The first three chapters in particular.) For a good discussion on the basics of chant in general: Gregori M. Sunyol, Introduccio a la Paleografia Musical Gregoriana. Montserrat: Abadia, 1925. For a better understanding of the relationship of sound patterning to oral transmission, see Thomas P.McCreesh, Biblical Sound and Sense: Poetic Sound Patterns in Proverbs 10-29. Sheffield: JSOT, 1991.

31 Music catalogues are still organized by Incipit.

32 The form and function of religious music -- and liturgies -- begin with Enheduanna back at Akkad. See, Wilson, Making the Hebrew Psalter.

33 The concept that scripts 'develop' is quite erroneous and stems from the conflation of a scribe with a calligrapher. Development implies that a letter change here, another there, until, finally, we have a new script. Scripts, however, are closed systems, carefully designed to work within the complex unity we call a writing system. Script families consist of a script, the class, and numerous mutations, fonts, descendants of the class. All modern "scripts," for example, are descendants, mutations, of precisely four script classes. All uppercase serifed fonts are descendants, mutations, of Roman Capitals and all lowercase serifed fonts are mutations of North African half-uncials; all modern sans-serif uppercase fonts are mutations of Roman Rustic Capitals and all lowercase fonts are mutations of Roman half-uncials. There are only two script classes for Hebrew, Paleo-Hebraic and Square Aramaic: The fonts used in the documents from, for instance, Gezer, are mutations of Paleo-Hebraic and the fonts used for the majority of the documents found in the Judean Desert and still used today are mutations of the Square Aramaic. Likewise, there are only two script classes for Greek, Attic Capitals and Constantine's ethnic-blend "Uncial." There are very few script classes in any writing system, no matter the language for which that script system is intended. 'New' script designs are extremely rare and occur under special -- and very predictable -- circumstances. There is no such thing as a "proto-typic" script; there must be a script class for a font to mutate from. <

Top of page