Zaima Annotation

1. Introduction

The goal of this document is to identify presentation and layout requirements of Zaima annotation and raise the questions that must be answered where the requirments are uncertain to then fully specify the presentation convention.

This document presents an overview of Ethiopic "Zaima" annotation and evaluates the suitability of available W3C standards for expressing the layout and presentation found in use cases. Characteristic samples have been selected that demonstrate the bredth and depth of the formatting complexity occuring in Zaima annotation. Recommendations are made for the addition or extension to an existing standard only when a mechanism is not available to support a requirement of Zaima presentation. The intention of this approach is to incur a minimal impact upon the existing family of applicable W3C standards and thus minimize the software modifications that would be required to support the annotation convention.

1.1 What is zaima?

Under divine inspiration in the early sixth century St. Yared (Miazia 5, 505 AD - Genbot 11, 571 AD) devised the system of over 600 notations (including 10 non-letter symbols) for his seminal work the "Diggua" (ድጓ) . Over the centuries that followed the disciples of St. Yared extended the annotation system to some 1140 notations. The notation indicates how the voice should rise and fall on a per-word and even per-letter basis in one of three named "moods". Zaima (ዜማ) literally means "church song, chant, hymn" of the kind recited during the Ge'ez liturgy of the Ethiopian Orthodox Church. The term is used here as a short hand for "YeKidus Yared YeZaima Miliktoch" (transcription for «የቅዱስ ያሬድ የዜማ ምልክቶች»), meaning "the religious music notation of Saint Yared". The convetion is also know as "Yaredic notation".

These moods are written in three distinct levels within the inter-linear space above the base text. In order from lowest to highest above the base text the moods are: Ge'ez (ግዕዝ), Izzel (ዕዝል), and Araray (ዓራራይ). The notation text is rendered in size aproximately a quarter to a third the scale of the base text. Izzel notation are written in red to help distinguish the mood from the other two (which in turn do not occur together in the absense of Izzel). When one mood does not exist for a given passage, the text of the upper levels may descend downward to occupy a lower level [Q2]. With the exception of the non-letter symbols (known as "ምልክት"), which may be employed by any of the three moods, the moods have their own distinct lexicons. Thus the mood context of an annotation row is also identifiable by the terms present. The annotation entities in turn will almost always appear in abbreviated form (known as "ሰራየ") to minimize their horizontal footprint in layout as well as to allow for faster rendering by calligraphers. Figure 1.1 shows an example over three lines.

Figure 1.1: A three line example of zaima notation.

1.2 The zaima box model

The W3C recommendations for Ruby annotation are directly applicable to Zaima practices. The CSS3 Ruby module applies a multi-level box model for describing ruby layout in the the writing practices of Asia. A modified box model is used hereafter to likewise describe Zaima presentation.

The simplified Zaima box model employs the same ruby base and adds two additional ruby text levels for a total of three. When all three moods exist for a passage their rows occur in sequence as shown in Figure 1.2:

Figure 1.2: Extended Ruby box model.

In practice the occurance of all three moods simultaneously is found only rarely. However, it is common that a single mood will occur twice, and occassionally three times, for a given passage. In these cases, the default, or primary, form of a mood will be at the lowest level [Q3]. The addition occurances of the same mood, considered alternatives, will appear at successively higher levels. The combination of moods and their alternatives maximally produces five levels of annotation text above the base. These scenarios will be illustrated in Section 2. A more comprehensive Zaima model is presented in Appendix A.

1.3 Axioms of zaima annotation

This section contains a collection of assertions for Zaima annotation that should be respected by any application implementing support for Zaima presentation.

[Q4] These rules must be vetted:

[Q4a] Izzel may not appear by itself.
[Q4b] Ge'ez may appear by itself.
[Q4c] Araray may appear by itself.
[Q4d] Ge'ez may combine with Izzel.
[Q4e] Araray may combine with Izzel.
[Q4f] Izzel must appear in red.
[Q4g] Ge'ez may not combine with Araray.
[Q4h] The Ge'ez row must appear beneath Izzel and Araray (when present).
[Q4i] The Izzel row must appear beneath Araray (when present).
[Q4j] The tonal marks (non-letter symbols) apply only to the letter below.
[Q4k] Annotations (abbreviated words, non-tonal marks) apply to the entire word below.
[Q4l] Tonal marks may be stacked together within a row for emphasis.
[Q4l] Annotations may not be stacked together.
[Q4m] No more than 3 tonal marks may occur in a stack.

2. Application of ruby markup to zaima use cases

W3C Ruby Annotation recomendation defines markup intended to accomodate the annotation practices of Asian scripts and offers a strong foundation for supporting Zaima annotation. In this section the W3C Ruby markup and CSS module style definitions are applied to increasingly complex samples of Ethiopic Zaima annotation. When a mechanism is not available from a W3C specification to support a specific Zaima use case, a mechanism is introduced and discussed. Notes are also made with every sample where the presentation rules are uncertain and when knowledge gaps are present.

2.1 Simple zaima

2.1.1 Single mood

In the simplest case of Zaima annotation only one mode occurs with annotation clearly associated with a single character. This simple case can be supported easily with the W3C Ruby markup:

<ruby>
  <rbc>
    <rb>አ</rb> 
    <rb>ሐ</rb> 
    <rb>ቲ</rb>
    <rb>፡</rb>
    <rb>እ</rb>
    <rb>ን</rb> 
    <rb>ተ</rb>
  </rbc>
  <rtc class="geez">
    <rt></rt>
    <rt>᎖</rt>
    <rt>ሳ</rt>
    <rt></rt>
    <rt>᎑</rt>
    <rt></rt>
    <rt>ዊ</rt>
  </rtc>
</ruby>
፡

Figure 2.1: Zaima annotation for a single mode (Selot Kidasie p.37).

While not essential for simple presentation cases, it is recommended that each rtc element be given an identifying class attribute. Clarifying the semantics of the mood with a class attribute will be an aid to validator tools that would check mode text for the proper use of the mode restricted lexicon. For example spell checking of the mood line and identifying invalid terms. Requiring a class attribute would assure that presentation rules are applied properly for the respective modes, aids machine understanding and transformation, and presentation in other target mediums such as text-to-speech browsers.

Recommended class values are the transcription names of the moods, sans punctuation: [ geez | izzel | araray ].

[class values of "ግዕዝ|ዕዝል|ዓራራይ" could be defined as aliases to their ascii counterparts]

	<ruby> <rbc> <rb>ባ</rb> </rbc> <rtc class="geez"> <rt>᎖᎓</rt> </rtc> <rtc class="geez" alt="1"> <rt>᎒᎒᎓</rt> </rtc> </ruby>

Figure 2.2: Zaima annotation for a single mode with one alternative (Selot Kidasie p.35).

In Figure 2.2 the attribute "alt" is introduced to indicate that the row is an "alternative" to the default level (zero level) of its class. It is strongly recommended that the alternative level be indicated explicitly by some means. The level should not be left to inference by the document order of the successive rtc of the same class type. Not all XML based languages are document order aware and semantics implied by order will be lost. For instance RDF is oblivious to element document order and transformations with technologies like GRDDL have been shown to be problematic (e.g. conversion of tables and arrays).

[Do any of the W3C recommendations already provide an index indicator that could be used in place of "alt"?]

[Q5] Are there rules that indicates when a class alternative should be used?]

Figure 2.3 presents an extreme case where as many as three alternatives may be given for a single mood:

<ruby>
  <rbc>
    <rb>ከ</rb>
    <rb>በ</rb>
    <rb>ሮ</rb>
  </rbc>
  <rtc class="geez">
    <rt>᎐</rt>
    <rt>᎐</rt>
    <rt>ር</rt>
  </rtc>
  <rtc class="geez" alt="1">
    <rt>᎖</rt>
    <rt></rt>
    <rt>ሥቱ</rt>
  </rtc>
  <rtc class="geez" alt="2">
    <rt></rt>
    <rt>᎓</rt>
    <rt>ሞኂ</rt>
  </rtc>
  <rtc class="geez" alt="3">
    <rt>በ᎖</rt>
  </rtc>
</ruby>

Figure 2.3: Zaima annotation for a single mode with three alternative (Selot Kidasie p.113).

Zaima annotation for a single mode with four alternative

<ruby>
  <rbc>
    <rb>ሊ</rb>
    <rb>ባ</rb>
    <rb>ኖ</rb>
    <rb>ስ</rb>
    <rb>ሃ</rb>
    <rb>፡</rb>
    <rb>ና</rb>
    <rb>ክ</rb>
    <rb>ር</rb>
    <rb>ር</rb>
  </rbc>
  <rtc class="geez">
    <rt rbspan="2"></rt>
    <rt rbspan="2">᎑ደረ᎓᎔᎖᎖</rt>
    <rt>᎖᎖</rt>
    <rt></rt>
    <rt>᎐</rt>
    <rt rbspan="2"></rt>
    <rt>ር</rt>
  </rtc>
  <rtc class="geez" alt="1">
    <rt>᎐</rt>
    <rt></rt>
    <rt rbspan="2">᎓᎔᎖᎓᎔᎓᎔᎖᎖</rt>
    <rt>᎖᎖</rt>
    <rt></rt>
    <rt>᎐</rt>
    <rt rbspan="2"></rt>
    <rt>ር</rt>
  </rtc>
  <rtc class="geez" alt="2">
    <rt>᎑</rt>
    <rt></rt>
    <rt>᎖መን</rt>
    <rt></rt>
    <rt>᎐</rt>
    <rt rbspan="3"></rt>
    <rt>ር</rt>
  </rtc>
  <rtc class="geez" alt="3">
    <rt rbspan="2"></rt>
    <rt>᎖᎔</rt>
    <rt rbspan="6"></rt>
    <rt>ር</rt>
  </rtc>
  <rtc class="geez" alt="4">
    <rt rbspan="2"></rt>
    <rt>᎒</rt>
    <rt rbspan="6"></rt>
    <rt>ር</rt>
  </rtc>
</ruby>፡

Questions:

[Q6] Is five rows the maximal case? If a limit is unknown, it should not be constrained.
[Q7] Verify the distribution over ባኖስ.

Figure 2.4: Zaima annotation for a single mode with four alternative (Zimarie Wemewasit p.34 (Seond Half)).

2.1.2 Two moods

Very often a passage can be recited in two moods. Valid co-occurances of moods are Ge'ez and Izzel, Araray and Izzel, and only rarely Ge'ez, Izzel and Araray. The Ge'z mood does not co-occur with Araray in the absense of Izzel. Figure 2.4 presents the simple case of two modes with no alternative enunciations:

<ruby>
  <rbc>
    <rb>ዬ</rb> 
    <rb>ዬ</rb> 
    <rb>ዬ</rb>
  </rbc>
  <rtc class="geez">
    <rt>ታ</rt>
    <rt>ፈየ᎖᎓</rt>
    <rt>ር</rt>
  </rtc>
  <rtc class="izzel">
    <rt>ሖ</rt>
    <rt>᎖᎓᎖᎓ጋሃ</rt>
    <rt>ር</rt>
  </rtc>
</ruby>
፡

Figure 2.5: Zaima annotation for two modes (Selot Kidasie p.177).

Figure 2.5 presents an extreme case where the Ge'ez mood occurs with two alternatives along with the Araray mood, creating a total of 4 interlinear levels:

Zaima annotation for two modes with 2 Ge'ez alternatives

<ruby>
  <rbc>
    <rb>ወ</rb> 
    <rb>ወ</rb> 
    <rb>ሀ</rb> 
    <rb>ብ</rb> 
    <rb>ከ</rb> 
  </rbc>
  <rtc class="geez">
    <rt>ቡር</rt>
    <rt></rt>
    <rt></rt>
    <rt></rt>
    <rt></rt>
  </rtc>
  <rtc class="geez" alt="1">
    <rt>᎖᎓᎔</rt>
    <rt>᎕</rt>
    <rt></rt>
    <rt></rt>
    <rt></rt>
  </rtc>
  <rtc class="geez" alt="2">
    <rt>ኢ</rt>
    <rt>ይ</rt>
    <rt>ኃ</rt>
    <rt></rt>
    <rt>ጐ᎖</rt>
  </rtc>
  <rtc class="izzel">
    <rt>᎙ሙላ</rt>
    <rt></rt>
    <rt>᎐</rt>
    <rt></rt>
    <rt>᎐᎖᎖</rt>
  </rtc>
</ruby>
፡

Questions:

[Q8] Verify that these are 4 rows and not a case of ᎖᎓᎔ stacking on top of ቡር (unlikely).
[Q9] Verify that Kurt is used here and not a smudged Rikrik.
[Q10] Kinet and Chiret glyphs vary more so than others, verify the correct use here.
[Q11] Verify that the bottom most Ge'ez mood is the primary. If it is not then we have a complication because the row index is not in step with the class index, and we need another attribute to indicate the row position because we do not want to rely on document order..

Figure 2.6: Zaima annotation for two modes with 2 Ge'ez alternatives (Merha Liqawnt p.127).

Figure 2.7 presents another extreme case where the Ge'ez mood along with Araray and its two alternatives for a total of 4 interlinear levels:

	<ruby> <rbc> <rb>ነ</rb> <rb>ጽ</rb> <rb>ር</rb> </rbc> <rtc class="geez"> <rt>᎑</rt> <rt>᎔᎓ራቀ</rt> <rt>᎒ዑ</rt> </rtc> <rtc class="izzel" alt="1"> <rt></rt> <rt></rt> <rt>᎐</rt> </rtc> <rtc class="izzel" alt="2"> <rt>᎖</rt> <rt></rt> <rt>ቡ</rt> </rtc> <rtc class="izzel"> <rt>᎖</rt> <rt>᎖᎓ራቀ</rt> <rt>᎖᎓᎙</rt> </rtc> </ruby> ።
Questions: [Q12] Verify that these are 4 rows and not a case of ቡ stacking on top of ᎐ (unlikely), and that the yizet is intended. [Q13] Verify that the top most Izzel is the primary. If it is then we have a complication because the row index is not in step with the class index, and we need another attribute to indicate the row position because we do not want to rely on document order..

Figure 2.7: Zaima annotation for two modes with 2 Izzel alternatives (Selot Kidasie p.253).

2.2 Complex zaima

2.2.1 Zaima text alignment and overhanging

Just as with Asian Ruby practices, annotation text may overhang the base text. This almost always occurs to the right side of a base letter, and rarely to the left. When a left side overhang is found it will be over the first letter of a word. Section 4.2 of the CSS3 Ruby Module defines the 'ruby-align' property for controling the alignment style of the ruby text with respect to the ruby base. The ruby-align property will be applied to Zaima use cases in this section to evaluate their suitability

Relatively rare case of Zaima with a left side overhang

፣<ruby>
  <rbc>
    <rb>ጕ</rb> 
    <rb>ር</rb> 
    <rb>ዔ</rb> 
    <rb>ሁ</rb> 
  </rbc>
  <rtc class="geez" style="ruby-align: right">
    <rt style="ruby-overhang: start">ውሡ</rt>
    <rt>ራ</rt>
    <rt>ሬ</rt>
    <rt style="ruby-align: left">ሃ</rt>
  </rtc>
  <rtc class="geez" alt="1" style="ruby-align: center">
    <rt style="ruby-overhang: start">ቡር᎓᎔᎓</rt>
    <rt></rt>
    <rt></rt>
    <rt></rt>
  </rtc>
  <rtc class="izzel" style="ruby-align: center">
    <rt style="ruby-overhang: start">᎙ሙ᎘</rt>
    <rt></rt>
    <rt>᎐</rt>
    <rt>᎐᎘</rt>
  </rtc>
</ruby>

`ruby-align: right`	`ruby-overhang: start`
`ruby-align: left`	`ruby-align: center`
CSS3 Ruby Module Figure 4.2.4, 4.2.5, 4.2.3, 4.3.2

Figure 2.8: Relatively rare case of Zaima with a left side overhang (Merha Liqawnt p.12).

More typically overhanging occurs to the right of a letter of base text and then usually at the end of a word as can be seen in the next two examples:

Zaima with right side distribute-space overhang

<ruby>
  <rbc>
    <rb>ኦ፡</rb> 
    <rb>መ</rb> 
    <rb>ዋ</rb> 
    <rb>ዒ</rb> 
    <rb>ት</rb> 
  </rbc>
  <rtc class="geez" style="ruby-align: right: ruby-overhang: start">
    <rt style="ruby-align: distribute-space; ruby-overhang: none">ጸ᎖᎓᎔᎖᎓᎔᎓᎔</rt>
    <rt></rt>
    <rt>ገብር</rt>
    <rt></rt>
    <rt>ኤል</rt>
  </rtc>
  <rtc class="izzel" style="ruby-align: center">
    <rt style="ruby-align:left; ruby-overhang: none">᎙ ᎙ ᎙</rt>
    <rt></rt>
    <rt>᎐</rt>
    <rt></rt>
    <rt>᎙</rt>
  </rtc>
</ruby>
፡

Questions:

[Q14] Check if <rb>ኦ፡</rb> is conceputally wrong. In principle punctuation is not annotated.

Figure 2.9: Zaima with right side distribute-space overhang (Merha Liqawnt p.30).

<ruby>
  <rbc>
    <rb>መ፡</rb> 
    <rb>ከመ</rb> 
  </rbc>
  <rtc class="geez" style="ruby-align: right: ruby-overhang: start">
    <rt style="ruby-align: distribute-space; ruby-overhang: none">ጸ᎖᎓᎔᎖᎓᎔᎓᎔</rt>
    <rt style="ruby-align: right">ወ</rt>
    <rt style="ruby-align: left">ተ</rt>
  </rtc>
  <rtc class="izzel" style="ruby-align: center">
    <rt style="ruby-align:left; ruby-overhang: none">᎙ ᎙ ᎙</rt>
    <rt></rt>
    <rt>᎐</rt>
  </rtc>
</ruby>፡

Questions:

[Q15] The difference in the separation between መ and ፡ in the base text, and the separation between ኦ and ፡ in the previous example is most likely accidental, and a standard distribution can likely be agreed to by expert caligraphers.
[Q16] Is it deliberate, and critical, that ወ hangs half way off the right side of መ?

Figure 2.10: Zaima with right side distribute-space overhang (Merha Liqawnt p.66).

Zaima with right side inter-word overhang

ሰ<ruby>
  <rbc>
    <rb>ብ</rb> 
    <rb>እ</rb> 
    <rb>፡</rb> 
    <rb>ሱ</rb> 
    <rb>ራ</rb> 
    <rb>ፌ</rb> 
  </rbc>
  <rtc class="geez">
    <rt>ለ</rt>
    <rt style="ruby-overhang: ???">ታ᎖᎓᎔᎖᎓᎔</rt>
    <rb></rb> 
    <rt style="ruby-align: right">ል</rt>
    <rb>ደ</rb> 
    <rb>ዊ</rb> 
  </rtc>
  <rtc class="izzel" style="ruby-align: center">
    <rt></rt>
    <rt style="ruby-overhang: start">᎐᎘᎙ሙ</rt>
    <rt></rt>
    <rt>᎘</rt>
    <rt>᎐</rt>
    <rt>᎐</rt>
  </rtc>
</ruby>ል፡

Questions:

[Q17] Verify that ለ is annotation for base letter ብ. In principle the abbreviated terms apply to the entire word base.
[Q18] Verify that ታ is annotation for እ (and not ብ). If so then we have both a left and right overhang occuring. [Perhaps a rule is needed here that a left single letter of left side overhang can occur (over left side punctuation or inter-letter space) in long annotations, while the rest hangs right.]
[Q19] Verify the markup of the Izzel mood.

Figure 2.11: Zaima with right side inter-word overhang (Merha Liqawnt p.64).

The above cases show overhangs at the end of word, extended across the Ethiopic wordspace. However it may occur between two letters within a word as well:

Zaima with right side overhang within a word

<ruby>
  <rbc>
    <rb>ኪ</rb> 
    <rb>ያ</rb> 
    <rb>ከ</rb> 
    <rb>፡</rb> 
    <rb>ኦ</rb> 
    <rb>ግ</rb> 
  </rbc>
  <rtc class="geez" style="ruby-align: right: ruby-overhang: start">
    <rt></rt> 
    <rt>ጽን</rt> 
    <rt></rt> 
    <rt style="ruby-align: distribute-space; ruby-overhang: start">ቡዘ᎖᎓᎔᎖᎓᎔᎓᎔</rt>
    <rt>መ</rt> 
  </rtc>
  <rtc class="izzel">
    <rt></rt>
    <rt>᎙</rt>
    <rt></rt>
    <rt></rt>
    <rt style="ruby-align:left; ruby-overhang: none">᎙ ᎙ ᎙</rt>
    <rt></rt>
  </rtc>
</ruby>፡

Questions:

[Q20] We want the annotation above ኦ to overhang only over the preceding ፡ and go no futher leftward.
[Perhaps a rule is needed here that left and right overhang, only overhangs a following punctuation or inter-letter space.]

Figure 2.12: Zaima with right side overhang within a word (Merha Liqawnt p.84).

Ruby annotation markup provides the rbspan attribute to distribute ruby text over ruby base letters. The need for the distribution of annotation text over two base letters is frequent in Zaima. The rbspan is applied in the next example:

<ruby>
  <rbc>
    <rb>ኪ</rb> 
    <rb>ያ</rb> 
    <rb>ከ</rb> 
  </rbc>
  <rtc class="geez">
    <rt>᎗</rt> 
    <rt>᎐</rt> 
    <rt>ቱ</rt> 
  </rtc>
  <rtc class="izzel">
    <rt>ረዩ</rt>
    <rt style="ruby-align: distribute-space" rbspan="2">᎓᎖᎓᎖᎓᎓᎖</rt>
  </rtc>
</ruby>፡

Figure 2.13: Zaima annotation distributed over two base letters (Merha Liqawnt p.103).

2.2.2 Annotation stacks

Very commonly the non-letter annotation symbols will be combined into a "stack" for emphasis. The Diret-Hidet tonal mark (᎘) is in fact a stacking of Diret (᎑) on top of Hidet (᎗). No other stackings will appear to fuse together as a single symbol and likewise no not receive their own Unicode codepoint. The exhaustive list of possible stackings is not known and software should not attempt to contrain stackings within the set of non-letter annoation symbols.

For the purpose of brevity in the following examples, a style class "stack" is defined for presentation of inline stacked items:

.stack { writing-mode:tb-lr; glyph-orientation-vertical: upright; }

	ዕ<ruby> <rbc> <rb>ጣ</rb> </rbc> <rtc class="geez"> <rt><span class="stack">᎑᎑</span></rt> </rtc> <rtc class="geez" alt="1"> <rt>᎓᎔᎓</rt> </rtc> </ruby>ን፡

Figure 2.14: Izzel Zaima with a Deret-Deret stack (Selot Kidasie p.87).

Finish Alignments
<ruby>
  <rbc>
    <rb>ሐ</rb> 
    <rb>ዋ</rb> 
    <rb>ር</rb> 
    <rb>ያ</rb> 
    <rb>ት</rb> 
  </rbc>
  <rtc class="geez">
    <rt style="ruby-align: center; ruby-overhang: right">᎑ሰማ</rt>
    <rt></rt>
    <rt></rt>
    <rt>ደ</rt>
    <rt style="ruby-align: left">ታት</rt>
  </rtc>
  <rtc class="izzel">
    <rt style="ruby-align: center; ruby-overhang: right">ሰማ᎓</rt>
    <rt></rt>
    <rt></rt>
    <rt rbspan="2" style="ruby-overhang: left">᎒᎖᎒᎓<span class="stack">᎔᎔<span>᎒</rt>
  </rtc>
</ruby>፡

Figure 2.15: Izzel Zaima with a Difat-Difat stack (Selot Kidasie p.129).

<ruby>
  <rbc>
    <rb>ፍ</rb> 
    <rb>ኖ</rb> 
    <rb>ት</rb> 
  </rbc>
  <rtc class="geez">
    <rt>አም</rt>
    <rt>ኑ<span class="stack">᎑᎔<span><span class="stack">᎑᎔<span></rt>
    <rt>ኃ</rt>
  </rtc>
  <rtc class="izzel">
    <rt></rt>
    <rt rbspan="3" style="ruby-align: center">᎙ሙ᎘</rt>
  </rtc>
</ruby>፡

Questions:

[Q21] Check if አምኑ was a single abbreviation and not two.

Figure 2.16: Ge'ez Zaima with a Deret-Difat stack (Merha Liqawnt p.22).

<ruby>
  <rbc>
    <rb>ዘ</rb> 
    <rb>ይ</rb> 
    <rb>ቀው</rb> 
    <rb>ም</rb> 
  </rbc>
  <rtc class="geez">
    <rt>᎐</rt>
    <rt>᎐</rt>
    <rt>᎒<span class="stack">᎔᎑<span>᎓᎖᎒᎓᎓</rt>
    <rt>ድድ</rt>
  </rtc>
</ruby>፡

Questions:

[Q22] Verify the distribution over ቀው.

Figure 2.17: Ge'ez Zaima with a Difat-Deret stack (Zimarie Wemewasit p.1).

<ruby>
  <rbc>
    <rb>ወ</rb> 
    <rb>ን</rb> 
    <rb>ጌ</rb> 
    <rb>ል፡</rb> 
    <rb></rb> 
    <rb>ቅ</rb> 
    <rb>ዱ</rb> 
    <rb>ስ</rb> 
  </rbc>
  <rtc class="geez">
    <rt>ወባ᎓᎖</rt>
    <rt>᎓᎖᎓᎓᎖</rt>
    <rt>᎓᎓᎖᎔</rt>
    <rt>᎓᎓᎖ዚኦዝ</rt>
    <rt></rt>
    <rt>ር</rt>
    <rt></rt>
  </rtc>
  <rtc class="izzel">
    <rt rbspan="2"></rt>
    <rt style="ruby-overhang: auto">᎓᎖᎓᎖᎒<span class="stack">᎔᎔᎔<span>᎒</rt>
    <rt rbspan="3"></rt>
    <rt>ነጽ</rt>
    <rt>᎗</rt>
  </rtc>
</ruby>፡

Questions:

[Q23] Verify the distribution over ል፡, should ል፡ be marked as a group or individuallyw or is it only ል that is marked with a right side overhang occurs over ፡?
[Q24] Check if the long ᎓᎖᎓... sequence is one contiguous annotation over the entire word, or many short annotations over each letter.

Figure 2.18: Izzel Zaima with stacks of three Difat (Selot Kidasie p.111).

Ge'ez Zaima with stacks of three entities

ወ<ruby>
  <rbc>
    <rb>ኮ</rb> 
    <rb>ነ</rb> 
  </rbc>
  <rtc class="geez">
    <rt>ብኪ</rt>
    <rt style="ruby-overhang: auto">᎓<span class="stack">᎔᎔᎔<span>᎓᎖</rt>
  </rtc>
  <rtc class="geez" alt="1">
    <rt style="ruby-overhang: auto"><span class="stack">᎑᎔᎔<span></rt>
    <rb></rb> 
  </rtc>
</ruby>፡

Questions:

[Q25] Verify the distribution over ነ.
[Q26] Verify that these are triple stacks, and not two rows where a double stack is on to of a single mark (or a single mark on top of a double stack).

Figure 2.19: Ge'ez Zaima with stacks of three entities (Zimarie Wemewasit p.68).

አ<ruby>
  <rbc>
    <rb>ሳ</rb> 
    <rb>ት</rb> 
    <rb>ኒ</rb> 
    <rb>፡</rb> 
    <rb>ማ</rb> 
    <rb>የ</rb> 
  </rbc>
  <rtc class="geez">
    <rt>ኪ</rt>
    <rt>ያሃ</rt>
    <rt>᎖᎔᎖᎔</rt>
    <rt>᎑<span class="stack">᎖᎖</span></rt>
    <rt>፴</rt>
    <rt>᎐</rt>
  </rtc>
  <rtc class="izzel">
    <rt>᎐</rt>
    <rt></rt>
    <rt rbspan="2" style="ruby-overhang: right">᎐ሙ᎘</rt>
    <rt style="ruby-align: right">᎐</rt>
    <rt></rt>
  </rtc>
</ruby>፡

Questions:

[Q27] Check if these are truly stacked Chirets.
[Q28] Check the distributions between ኒ and ማ for both Ge'ez and Izzel.

Figure 2.20: Zaima with possible stackings of Chiret (Merha Liqawnt p.19).

	ዓ<ruby> <rbc> <rb>ለ</rb> <rb>ም</rb> </rbc> <rtc class="geez"> <rt><span class="stack">᎖᎖<span></rt> <rt style="ruby-align: right">ሉ᎙</rt> </rtc> </ruby>፡

Questions: [Q29] Check if these are truly stacked Chirets. [Q30] If the Chirets are not stacked, is the over-lapping alignment intentional or accidental?

Figure 2.21: Zaima with a (possible?) stack of two Chirets (YeZaima Arist Milikt p.25).

	<ruby> <rbc> <rb>አ</rb> <rb>ን</rb> <rb>ቲ</rb> </rbc> <rtc class="geez" style="ruby-align: left"> <rt>᎔᎓</rt> <rt rbspan="2"><span class="stack">᎖᎖᎖<span></rt> </rtc> </ruby>፡

Questions: [Q31] Check if these are truly stacked Chirets. [Q32] Check if left side Chiret alignment is intentional or accidental.

Figure 2.22: Zaima with a (possible?) stack of three Chirets (YeZaima Arist Milikt p.25).

2.2.3 Mixed moods

In some instances mood text may appear distributed across several presentation rows, such that two moods may share a single row of presentation. This may be an artifact of edits made at various times that prevented the mood text from appearing on the same line of presentation.

[Q33]: General questions for the mixed mood samples:

[Q33a] When Izzel text appears in several rows, are they indepented annotations, or part of the same?
[Q33b] If we are seeing one sequence of Izzel Zaima, but spread over several presentation rows, is anything lost if the same Izzel were presented on a single row?
[Q33c] Is there any significance to Izzel appearing below Ge'ez or above Araray? Likewise when Ge'ez and Araray change their presentation row?


Questions: [Q34] How many rows of Izzel and Ge'ez (or Araray) are in this sample?

Figure 2.23: Zaima with Izzel appearing in 3 levels (Selot Kidasie p.145).


Questions: [Q35] How many rows of Izzel are in this sample?

Figure 2.24: Zaima with one Ge'ez mood and three in Izzel (Selot Kidasie p.157).


Questions: [Q36] Is the Izzel and Ge'ez (or Araray) Zaima on the same line done for convenience, or is it meaningful? [Q37] How many rows of Izzel and Ge'ez (or Araray) are in this sample?

Figure 2.25: Ge'ez Zaima with Izzel -and Araray? (Selot Kidasie p.189).


Questions: [Q38] Is the Izzel text on its own presentation row? If so, is it required that the presentation line appear vertically between two others?

Figure 2.26: Three Ge'ez Zaima levels with Izzel (Selot Kidasie p.213).


Questions: [Q39] Are the three Chirets on a presentation row half-way between the two Ge'ez rows?

Figure 2.27: Zaima with two Ge'ez moods and three in Izzel (Selot Kidasie p.221).


Questions: [Q40] How many rows of Izzel are in this sample?

Figure 2.28: Zaima with Izzel appearing in 3 levels (Selot Kidasie p.263).


Questions: [Q41] Verify that this 3 rows of Izzel and 3 rows of Ge'ez (and/or Ararary) are presented.

Figure 2.29: Zaima with three Ge'ez moods and three Izzel moods distributed over four rows (Zimarie Wemewasit p.68).

2.2.4 Lesser understood formatting

Similar to the mixed mood presentation in Section 2.2.3, other presentation phenomena can be found that may simply be the result of later date edits, or some caligraphic artifact.


Questions: [Q42] What is the correct distribution of ርእሰዊ over ተ፡ሊ? [Q43] Is hulet neteb (፡) in the base considered marked by the Zaima? [Q44] Is ርእ, ሰ and ዊ on seperate rows, or are they one group that rises vertically (and is rising vertically deliberate)?

Figure 2.30: Zaima rising from Ethiopic wordspace (Selot Kidasie p.219).


Questions: [Q45] What is the correct distribution of እግዚ over ሌ፡ወ? [Q46] Is hulet neteb (፡) in the base considered marked by the Zaima? [Q47] Is እግዚ rising vertically to the right deliberate?