Linguistic Divergence in Armenian Bible and Lectionary Palimpsests

Full text

Open Access. © 2025 the author, published by De Gruyter. This work is licensed under the Creative
Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
https://doi.org/10.1515/9783111552170-012
Hasmik Sargsyan
Linguistic Divergence in Armenian Bible and
Lectionary Palimpsests
Abstract: This paper’s subject is variant readings of linguistic relevance in Old
Armenian Bible and lectionary fragments preserved as lower layers of palimp-
sests. The aim is to contribute to the linguistic research on these rarely or never
studied witnesses of the Old Armenian language. Such research will enhance
understanding of the linguistic diversity of Old Armenian and, as may be the case,
give us linguistic tools to determine the time and place at which the textual wit-
nesses were produced. To these ends, the paper presents the first results of re-
search into variant readings of linguistic importance in several Bible and lection-
ary fragments.
 Research on palimpsest fragments
. The relevance of palimpsest fragments for linguistic
variation in Old Armenian
The study of palimpsests containing fragments of Armenian texts as their lower
layers is essential for several reasons. Only a small number of studies have dealt
with the linguistic divergence and variety of Old Armenian, and this is true for
both the diachronic and synchronic perspectives. The diachronic perspective
relates to the changes observed in the Old Armenian language over time, as evi-
denced by the written documents preserved from the inception of Armenian lit-
eracy to the present day. The synchronic perspective concerns the linguistic varie-
ty of Old Armenian as reflected in written texts that date from the time of the
emergence of Armenian literacy in the fifth century. Up to the present day, our
understanding of the linguistic characteristics of Old Armenian, notably during
the period from the fifth to the tenth century, remains limited from both these
perspectives.1

1 The existing studies on the language and orthography of the first centuries of Armenian litera-
cy include Meillet 1903; Acharyan 1928; Gyulbudaghyan 1973; Weitenberg 1994; Weitenberg 1997;
Weitenberg 2006; and Künzle 1984. The language of Old Armenian texts from the earliest stages of
its literacy appears quite uniform; however, researchers have pointed out linguistic features that

  Hasmik Sargsyan
The scarcity of studies on the linguistic divergence of Old Armenian is in part
due to the limited number of textual witnesses (manuscripts and inscriptions)
dating back to the first centuries of Armenian literacy. The oldest dated manu-
scripts of Armenian go back to the ninth century, and fewer than twenty manu-
scripts are dated before the year 1000.2 Palimpsest fragments are of special im-
portance in the study of the earliest stages of the Armenian language since usually
they can be assigned to an older period than the oldest dated non-palimpsest
manuscripts preserving the same texts. The groundwork for studying palimpsests
containing Armenian undertexts was established through the edition of the Ar-
menian layer of two palimpsests with Georgian overtexts – namely, Sinai, St Cath-
erine’s Monastery, georg. NF 13 and 55 (hereafter: Sin. georg. NF 13 and NF 55) –
which also contain the only known manuscript witnesses of the Caucasian Alba-
nian language in their lower layer.3 The study of palimpsest fragments is relevant
for understanding the history of text transmission, the development of written
tradition, and the linguistic variation of Armenian during the early centuries of its
literacy.4 And, as noted by Antoine Meillet, the precise study of the Old Armenian
language must be conducted based on the oldest preserved witnesses of Armenian
literature.5 Furthermore, it is likely that the few existing textual witnesses of the
first centuries of Armenian literacy represent not the natural state of spoken Old
Armenian of the time but only a written standard of it, and all we can hope for are
mere indications of change and variety within Old Armenian. For this reason,
every significant case of divergence in the textual tradition must be considered.
To date, scholarly literature predominantly focuses on the Classical, Modern
Eastern, and Modern Western Armenian standards (and occasionally Middle Ar-

suggest the potential existence of other varieties alongside the written standard (see, for example,
Weitenberg 1996 and 2014).
2 See, for example, Kouymjyan 2015, 38. Beda Künzle (1984) has used two of the oldest manu-
scripts for his edition of the Gospels: Yerevan, Matenadaran, M 6200, of 887 CE (formerly ms 1111
of the library of the Lazarev Institute of Oriental Languages, Moscow) and M 2374, of 989 CE (for-
merly Ejmiatsin, Armenian Patriarchate, 229).
3 See Gippert 2010 for the edition of the Armenian undertexts, Gippert et al. 2008 on the Albani-
an palimpsests, and Gippert 2022 for the perspectives of working on Armenian palimpsests in
general.
4 See Stone 1993 for the assessment of linguistic variants and Weitenberg 2014 on the methodol-
ogy of determining dialectal variation in Armenian manuscripts.
5 Meillet 1903, 489: ‘C’est sur l’observation de la graphie attestée par l’accord des vieux manu-
scrits demeurés fidèles à l’original que doit reposer une étude précise de l’ancienne langue armé-
nienne’ (‘A precise study of the Old Armenian language must rely on the examination of the
spelling attested by common accord of the ancient manuscripts that have remained faithful to the
original’).

Linguistic Divergence in Armenian Bible and Lectionary Palimpsests  
menian), with limited descriptions of dialectal variation in Armenian, as the main
concentration is on modern dialects. Critical editions of texts and dictionaries
attributed to one of the mentioned strata are adapted to the known grammatical
and orthographic rules of the corresponding standard. The editions of biblical
texts, too, often tend to emend linguistic and orthographic deviations from the
Classical Armenian standard, regarding them as ‘erroneous’, especially when the
edition is based on a large number of manuscripts.6 However, as Michael Stone
has pointed out, it is not the task of critical editions to reflect all the variations
occurring in every manuscript,7 as that would be simply impossible.
Even considering that there are probably more studies on the Bible than on
other texts written in Old Armenian, many linguistic aspects of the study of the
Armenian biblical texts remain open. Hence, the study of manuscript- and text-
specific linguistic divergences is crucial for the study of the Armenian language in
general and for the earlier stages of Old Armenian in particular.
. Palimpsest fragments within DeLiCaTe
Within the framework of the Development of Literacy in the Caucasian Territories
(DeLiCaTe) project,8 research on the linguistic divergence documented in Armeni-
an palimpsests is a work in progress. The following are the palimpsest fragments
that the project has dealt with so far. Apart from Sin. georg. NF 13 and NF 55,
which contain fragments of the Pauline Epistles with the Euthalian apparatus and
of the Old Testament books attributed to King Solomon,9 another palimpsest with
an Armenian underlayer is preserved at St Catherine’s Monastery: syr. 7, with
fragments from the Epistle to the Hebrews in the undertext of its fols 47 and 48.10

6 This is true of, e.g., Zōhrapean 1805; Amalyan 1996; and Alexanian 2012. Yovhannēs Zōhrapean
does provide variant readings of the manuscripts he has used, but he does not note which variant
readings are contained in which manuscripts. Hayk Amalyan’s (1996, 27–28) objective is to recon-
struct the original form of the translation of the Bible into Armenian by comparing existing
manuscripts and correcting any perceived ‘errors’ within them. He also points to the fact that
Zōhrapean has kept all the ‘errors’ of his manuscripts, in an endeavour to render them as faith-
fully as possible. Joseph M. Alexanian (2012, xi) states that the focus of his edition is ‘on a text and
apparatus that may reflect differences in Syriac and Greek Vorlage, rather than on a presentation
of all manuscript variants to illustrate the development of the Armenian language’.
7 Stone 1993, 24.
8 For details, see Jost Gippert’s contribution to the present volume and the project website at
<https://www.csmc.uni-hamburg.de/research/current-projects/delicate.html> (accessed on 9 April 2024).
9 Gippert 2010.
10 Hebrews 10:20–35, 11:25–39; see also Gippert 2023, 214.

  Hasmik Sargsyan
Palimpsests from other collections include Birmingham, Cadbury Research Li-
brary, Mingana Collection, Christ. Arab. Add. 124, which also contains fragments
from Hebrews.11 Palimpsests of the Mesrop Mashtots Institute of Ancient Manu-
scripts (Matenadaran) in Yerevan include
− Fragment no. 35 (John 7:44–52)
− M 196 (Proverbs 2:11–18, 2:4–11)
− Fragment no. 461 (Luke 4:8–11)
− M 470 (Acts 25:8–14, 26:14–20)
− M 963 (Luke 24:51–53)
− M 3850 (Acts 15:38–41, 16:1–4).
Lectionary fragments are preserved in M 1306, M 2166, and M 4435 (see Section 3.2
below).
. Palimpsested Bible fragments vs lectionary fragments
There are two important issues that one must keep in mind while dealing with
palimpsested Bible and lectionary fragments. The first is their dating: in most
cases, the only reliable basis is the dating of the upper layer (if there is any), yield-
ing a terminus ante quem. A more precise dating is usually not possible, at least
not until well-defined and sound palaeographical, linguistic, and historical indica-
tions, among others, have been taken into consideration. Determining the most
significant indicator(s) for dating the lower layer of a palimpsest heavily relies on
factors such as the history of the relevant language(s) and script(s), the history of
the given texts and manuscripts, and various other related considerations. The
second challenge lies in distinguishing between palimpsested Bible fragments
proper and lectionary fragments. Lectionaries primarily feature readings from
the Bible arranged according to the liturgical calendar. This similar content makes
it difficult to consistently and clearly differentiate between these two types of
texts within palimpsest fragments. In contrast to Bible manuscripts proper, identi-
fiable structural elements of lectionaries are, firstly, the titles of liturgical units,
along with the non-sequential rendering of biblical passages from the Old and
New Testaments. In this paper, fragments that could not be further identified as
part of a lectionary due to the absence of structural clues have been provisionally
treated as Bible fragments.

11 Hebrews 11:14–33; see also Gippert 2023, 214.

Linguistic Divergence in Armenian Bible and Lectionary Palimpsests  
 Orthographic variation in Old Armenian and its
assessment
. Orthographic variation in Old Armenian
The textual transmission of the Armenian Bible is comparatively consistent. However,
the Armenian manuscripts, and especially palimpsest fragments, do show some varia-
tion. Till today, we have few descriptions of the orthographic and language variation
in the Old Armenian manuscripts. They include Antoine Meillet’s short account on the
writing style (French graphie) of the Armenian Bible, Beda Künzle’s description of the
variation in orthography and linguistic expression in the Gospel manuscripts E and M
(see Section 2.3 below), as well as several accounts by Jos J. S. Weitenberg.12
Orthographic peculiarities of a given written specimen are in the first place con-
nected with the phonetic characteristics behind them. Hrachya Acharyan, Sirak
Gyulbudaghyan,13 and most other linguists dealing with the topic agree that Armenian
orthography from the first centuries after the creation of the alphabet was phonetical-
ly based. This means that differences in orthography reflected differences in phonet-
ics, that is, pronunciation. Discussing the phonetic values of the Armenian letters and
their combinations, Heinrich Hübschmann14 and, following him, Acharyan point to a
few tools for determining the possible pronunciation of the written evidence. These
tools include the arrangement of the letters in the Armenian alphabet (in its compari-
son with the Greek alphabet); the Armenian version of the Grammar by Dionysius
Thrax; foreign words transliterated into Armenian; and Armenian words transliterat-
ed into other languages.15 Gyulbudaghyan, who offers an account on the orthography

12 Meillet 1903; Künzle 1984, 58*–85*; Weitenberg 1994; Weitenberg 1997; Weitenberg 2006.
13 Acharyan 1928, 286–346; Gyulbudaghyan 1973, 63.
14 Hübschmann 1876, 60.
15 Hübschmann 1876, 60: ‘Diese [the pronunciation of Old Armenian] zu bestimmen haben wir drei
Hilfsmittel, 1) die armenische Bearbeitung des Dionysius Thrax, die jedenfalls alt ist, wenn sie auch
nicht, wie angenommen wird, in das 5. Jahrhundert hinaufreichen sollte, 2) das armenische Alphabet,
3) die Transcriptionen aus und in das Armenische’. Acharyan 1928, 289: ‘Հայերէն այբուբենի Ե
դարու կամ ընդհանրապէս խօսելով՝ հին հնչումը որոշելու համար չորս միջոց կար ասոնք
են՝ 1. Հայերէն այբուբենի դասաւորութիւնը։ 2. Դիոնեսիոս Թրակացւոյն քերականութիւնը։
3. Օտար լեզուներէ տառադարձուած բառերը՝ հայերէնի մէջ։ 4. Հայերէնէ տառադարձուած
բառերը՝ օտար լեզուներու մէջ’ (‘There are four means for determining the pronunciation of the
Armenian alphabet in the fifth century or, roughly speaking, the old pronunciation of Armenian.
These are: 1. the order of the letters in the Armenian alphabet; 2. the Grammar by Dionysius Thrax; 3.
the spelling of words transliterated from other languages into Armenian; 4. the spelling of words
transliterated from Armenian into other languages’).

  Hasmik Sargsyan
of the fifth- to ninth-century texts,16 emphasises the problem of determining the or-
thography of Armenian from the fifth to sixth centuries solely based on ninth-century
manuscripts. Therefore, he uses epigraphic material and evidence from loanwords.
He also indicates ‘certain orthographic deviations’ as a possible indicator.17
. ‘Linguistic variants’
When searching for linguistic features that can or might indicate an older stage
than the oldest dated manuscript witnesses of the Old Armenian language, the
first step is to look for non-coincidental patterns of divergence in orthography and
language. For this paper, only critical editions of the Bible and lectionaries were
taken into consideration. It is clear to me that a more comprehensive analysis
must additionally include a comparison with the manuscripts themselves, since
the critical editions often harmonise orthographic and other deviations.
We must certainly bear in mind that not all variant readings indicate a
change or variation in a linguistic sense, no matter whether the scribe inserted
them unknowingly or deliberately.18 As some researchers have suggested,19 the
assessment of ‘linguistic’ variants, as coined by Michael Stone,20 requires a sys-
tematic collection and description of variant readings in Armenian manuscripts of
the fifth to eleventh centuries outside the critical apparatus, as no critical edition
can satisfy every researcher’s needs. The next step involves situating these sys-
tematic descriptions within a broader dialectal and chronological context.21 As
mentioned by Stone, cooperation between textual critics and linguists is crucial in
this respect.22 Such cooperation is one of the goals of the DeLiCaTe project.

16 Gyulbudaghyan 1973, 12–65.
17 Gyulbudaghyan 1973, 12: ‘Դժվար է Թ դարի ձեռագրերով որոշել 5–6-րդ դարերի
ուղղագրությունը, բայց մեզ օգնում են վիմական արձանագրությունները, որոնք թեև ծավալով
փոքր են, բայց հնագույն շրջանից են մնացել․ օգնում են նաև հայերենից օտար լեզուներին և
օտար լեզուներից հայերենին անցած բառերի, փոխառությունների տառադարձությունը,
ինչպես և ուղղագրական առանձին շեղումները’ (‘Deciding on the orthography of the fifth to
sixth centuries based on ninth-century manuscripts is challenging. However, inscriptions come to
our aid, despite their brevity, as they originate from ancient times. The spelling of Armenian
words transliterated into other languages and vice versa, along with certain orthographic devia-
tions, also proves to be useful’).
18 Stone 1993, 20.
19 Weitenberg 2014, 223; Stone 1993, 20.
20 Stone 1993, 20.
21 Weitenberg 2014, 223.
22 Stone 1993, 21.

Linguistic Divergence in Armenian Bible and Lectionary Palimpsests  
. First case studies based on palimpsest fragments
In the following sections, I discuss some orthographic and linguistic variations
that can be found in the palimpsested Armenian Bible and lectionary fragments
listed above. The following issues are dealt with: some instances of the ortho-
graphical differentiation between <e> and <ē>,23 and the simplification of the
consonant cluster <rh> → <h>.
Two critical editions were consulted for the comparison with the Bible and
lectionary fragments preserved as the lower layers of the palimpsests: the com-
plete edition of the Old and New Testaments of 1805 by Yovhannēs Zōhrapean
(also called the Zohrab Bible), based on several manuscripts of the Mechitarist
collection in Venice, the oldest of which dates from the fourteenth century, and
Künzle’s edition of 1984, which renders two of the oldest and most complete man-
uscripts preserving the Armenian Gospels, namely, M = Yerevan, Matenadaran,
M 6200 (formerly kept in the Lazarev Institute of Oriental Languages in Moscow),
dated 887, and E = Yerevan, Matenadaran, M 2374 (formerly housed in the Arme-
nian Patriarchate in Ejmiatsin), dated 989.
 Orthographic variation between <e> and
<ē> in palimpsest fragments
. The distribution of ե <e> and է <ē> in Classical Armenian
Due to the presence of two distinct graphemes in the Armenian alphabet created by
Mesrop Mashtots for what today is an /e/-sound in both Modern Armenian standard
languages (East and West Armenian), we may assume that ե <e> and է <ē> initially
represented two different sounds.24 Etymologically, է <ē> goes back to the diphthong
[ey]. Manuk Abeghyan assumes that ե <e> was pronounced as an open vowel and է

23 See Künzle 1984, 61* and 63* for a short account as well as Weitenberg 2006 ‘on the use of
Armenian e and ē to render Greek ε and η in onomastic material in the Gospels and in the book of
Deuteronomy’ (Weitenberg 2006, 215).
24 See Schmitt (1972) for the transliteration of Armenian, and specifically page 300 for the repre-
sentation of է as <ē>. It is worth noting that the transliteration convention, using what appears to
be a long vowel, <ē>, to denote Է and implying a distinct pronunciation from ե <e>, is a practice
going back to the works of Heinrich Hübschmann, Antoine Meillet, and Émile Benveniste. Addi-
tional details can be found in Schmitt 1972, 297; see also Hübschmann 1882, 31–39, and Hübsch-
mann 1895–1897, 1–2.

  Hasmik Sargsyan
<ē> as a closed one.25 The same author points to the appearance of ե <e> in the con-
junction եթե <etՙe>, թե <tՙe> (‘if, that’, etc.), in the past tense of the copula (եի <ei>
first-person singular, եիր <eir> second-person singular, եաք <eakՙ> first-person
plural, etc.), and in the imperfect (գրեիք <greikՙ> second-person plural of the verb
‘write’, etc.), which ‘later became’ էի <ēi>, էիր <ēir>, էաք <ēakՙ>, and so on.26
Gyulbudaghyan adduces instances in which ե <e> is found in documents of the
fifth to ninth centuries instead of the standard writing with է <ē>.27 For this purpose,
he considers epigraphic material.28 He outlines two positions in the interpretation of
this variation, as represented by Antoine Meillet and Eduard Aghayan.29 Gyulbu-
daghyan disagrees with Meillet’s approach, which suggests that the orthography of
the imperfect with ե <e> is older than that with է <ē>. Rather, he aligns with
Aghayan’s viewpoint, according to which the spellings with ե <e> and է <ē> are mere-
ly confused in the textual witnesses.30 The question remains whether the deviations
from the known (or, in Gyulbusaghyan’s wording, ‘accepted’) rules concerning the
distribution of ե <e> and է <ē> in Old Armenian themselves follow a rule-based pat-
tern or are entirely at random. The answer to this question might be different de-
pending on the time and place at which the written materials were produced (and
perhaps also on the background of a given scribe). In other words, the question is
whether a chronological or geographical pattern can be established.
Discussing the divergence in several Gospel manuscripts from Ejmiatsin and the
Moscow Gospels (M), Meillet points to the fact that some of these manuscripts (five in
total, ‘les manuscrits corrects’) show consistency in many points, including the or-
thography of the imperfect and of թե <tՙe>, եթե <etՙe> (‘if, that’) with ե <e>. However,
three of those manuscripts have undergone more influence of ‘the Armenian of the
Middle Ages’ (‘l’arménien au moyen âge’) and show a less consistent orthography.31
Meillet’s statement can be understood in the following way: if certain manuscripts

25 Abeghyan 1936, 9.
26 Abeghyan 1936, 13: ‘բայերի անցյալ անկատարը, վոր հնագույն գրությամբ յեղել ե՝ եի,
եիր, եաք, եիք, եին գրեի, գրեիր, գրեաք, գրեիք, գրեին, հետագայում դարձել ե՝ էի, էիր, էաք
ևայլն’ (‘the past imperfect of the verbs that was written with ե <e> according to the old writing,
եի <ei> ‘(I) was’, եիր <eir> ‘(you, singular) were’, եաք <eakՙ> ‘(we) were’, եիք <eikՙ> ‘(you, plural)
were’, եին <ein> ‘(they) were’, գրեի <grei> ‘(I) wrote’, գրեիր <greir> ‘(you, singular) wrote’,
գրեաք <greakՙ> ‘(we) wrote’, գրեիք <greikՙ> ‘(you, plural) wrote’, գրեին <grein> ‘(they) wrote’,
later became էի <ēi>, էիր <ēir>, էաք <ēikՙ>, and so on’).
27 Gyulbudaghyan 1973, 40–45.
28 Gyulbudaghyan 1973, 40–41.
29 Gyulbudaghyan 1973, 42.
30 Meillet 1903, 491; Aghayan 1964, 358–359.
31 Meillet 1903, 490–491.

Linguistic Divergence in Armenian Bible and Lectionary Palimpsests  
consistently use a particular spelling, there must be a reason for that consistency.
Specifically, one must infer that the manuscripts demonstrating consistency are more
closely associated with a time when the pronunciation of ե <e> differed from that of է
<ē>, therefore possibly indicating an older form. After the differences in the pronun-
ciation of the two letters were neutralised, the orthography also changed and con-
ceivably became less consistent over time.
In the palimpsest fragments dealt with in the DeLiCaTe project (see Section 1.2),
the conjunctions թե <tՙe> and եթե <etՙe> (‘if, that’ etc.) are always written with ե <e>;
the same consistency can be observed in the spelling of the imperfect with ե <e>.
This likely suggests that the orthography with ե <e> is older than that with է <ē>.
In Sections 3.2 and 3.3, I introduce a few specific instances of the orthographic
variation between ե <e> and է <ē>.
. M 4435
The manuscript Yerevan, Matenadaran, M 4435 is a Gospel codex dated to the year
1424, to which flyleaves with erased text were bound. According to the Matena-
daran catalogue, the flyleaves’ undertext is also from the Gospels, written in
round erkatՙagir majuscules, and they even comprise a miniature.32
The four pages of the flyleaves represent one original folio that was cut in
two, with the resulting page order 3 + 6 (Fig. 1a) and 4 + 5 (Fig. 1b).33 The text is
arranged in two columns, with 9 + 9 lines and 17–21 characters per line in each
column. One or two characters are missing in the left column of pages 3–6. The
flyleaves were erased but are not overwritten. In contrast to the catalogue de-
scription, the undertext of M 4435 contains a lectionary fragment with readings
from the Old Testament (and not from the Gospels), intended to be recited during
the ninth hour of Palm Sunday and on Easter Monday and featuring a reading
from Genesis (1:1–7) and Psalm 117:26–27.34 On page 5 (Fig. 1b), one can observe the
miniature which is also mentioned in the catalogue.

32 Yeganyan, Zeytunyan and Antabyan 1965, 213; see the Introduction by Jost Gippert to this
volume for more information on the miniature.
33 In the case of flyleaves, pages are numbered instead of folios in the catalogues of the Matenadaran.
34 See Renoux 1971, 258–261 [120–123] for the corresponding passage in the Jerusalem-rite lec-
tionaries.

  Hasmik Sargsyan
Fig. 1a: Yerevan, Matenadaran, M 4435: p. 3 (top) + p. 6 (bottom); red-cyan pseudocolour image,
© Mesrop Mashtots Institute of Ancient Manuscripts (Matenadaran), Yerevan / DeLiCaTe project.

Linguistic Divergence in Armenian Bible and Lectionary Palimpsests  
Fig. 1b: Yerevan, Matenadaran, M 4435: p. 4 (top) + p. 5 with miniature (bottom); red-cyan pseudocolour
image, © Mesrop Mashtots Institute of Ancient Manuscripts (Matenadaran), Yerevan / DeLiCaTe project.

  Hasmik Sargsyan
Page 6 of M 4435 contains an instance of the imperfect form <šrǰer>
instead of the expected <šrǰēr> (‘(it) was hovering’), see example (1) from
Genesis 1:2:
(1) M 4435, p. 6, col. a, ll. 12–13
<>{ } [ ] < >{ }
<e>{w} [h]ogi a y35 šrǰer i <v>{e}ray ǰowrcՙ
‘And the Spirit of God was hovering over the waters’
. M 3850
Another attestation of an ե <e> for an expected է <ē> is provided by <hrei>
(vs <hrēi>) (‘of a Jew(ess)’) in Acts 16:1. This verse is preserved on a palimp-
sested flyleaf (page 6, Fig. 2) of Yerevan, Matenadaran, M 3850, a fifteenth-century
Gospel codex.36 The flyleaf belongs together with page 5; the original folio was cut
in two and additionally clipped at one edge (the left side of page 5 and the right
side of page 6), with only the upper half of the folio being preserved.
Fig. 2: Yerevan, Matenadaran, M 3850: p. 5 (left) + p. 6 (right); red-cyan pseudocolour image,
© Mesrop Mashtots Institute of Ancient Manuscripts (Matenadaran), Yerevan / DeLiCaTe project.
The erased text, written in round erkatՙagir majuscules in two columns, contains
Acts 15:38–41 and 16:1–4, with seven (of originally fourteen) lines preserved in each

35 We cannot be certain if the genitive of <astowac> (‘God’) cited here is an abbre-
viated form of <astowcoy> or <astowacoy>, until a written attesta-
tion of its full form is found in the oldest Armenian manuscripts.
36 Yeganyan, Zeytunyan and Antabyan 1965, 1096–1097.

Linguistic Divergence in Armenian Bible and Lectionary Palimpsests  
column; there are 8–10 characters in each line. <hrei> appears in Acts 16:1 on
page 6:
(2) M 3850, p. 6, col. a, ll. 1–6
[ ] [ ] [ ] [ ] [ ]
[ ]
ašakert omn ēr anown timovtՙeos ordi knoǰ hrei hawatacՙeloy
‘a certain disciple was there, named Timotheus, the son of a believing wom-
an, a Jewess’
. Further considerations
If the usage of ե <e> predates that of է <ē> in examples (1) and (2), one must also
consider why and how է <ē> came to replace ե <e> in specific contexts. It is worth
noting that է <ē> seems to occur less frequently in Armenian orthography as com-
pared to ե <e>. Aghayan reinforces his argument that the use of է <ē> in the im-
perfect is the ‘correct’ Mesropian orthography by pointing out that, in imperfect
forms, է <ē> (or ե <e>, depending on which one was in the Mesropian orthogra-
phy) consistently precedes a vowel.37 In Aghayan’s perspective, this suggests that
between the է <ē> (or ե <e>) and the subsequent vowel, a -y must have been pro-
nounced but not written, given that էա <ēa> (or եա <ea>) was not a diphthong in
the imperfect. Whether եա <ea> or էա <ēa> were pronounced as diphthongs in
the imperfect by the time the alphabet was created is hard to determine. Howev-
er, the fact that ե <e> or է <ē> precede vowels in the imperfect might be a plausi-
ble reason for the confusion and the shift in orthography from ե <e> to է <ē> in
the contexts given above. Etymologically, the imperfect of Armenian goes back to
a formation of the type Arm. bere- < Proto-Indo-European *bhere- and the Armeni-
an innovation for the first-person singular suffix of the past, -i.38 The proximity of
ե <e> to ի <i> could imply that the pronunciation of the ե <e> was affected by the
following vowel. Beyond a chronological differentiation, a geographical distribu-
tion of the pronunciation of ե <e> vs է <ē> in the early centuries of literacy cannot
be ruled out.

37 Aghayan 1964, 359.
38 Klingenschmitt 1982, 14–15.

  Hasmik Sargsyan
 Consonant cluster simplification <h> vs <rh>
Yerevan, Matenadaran, M 196 is a codex of the thirteenth century containing
books from the New Testament (Acts and Catholic Epistles). The two palimpsested
pages 60d and 65a (sic)39 of M 196 are written in round erkatՙagir majuscules, ac-
cording to the catalogue. In their reconstructed order 65a + 60d (see Fig. 3a), the
two pages were presumably originally two columns of the same page, turned by
180°. Their undertext contains Proverbs 2:4–18.
Fig. 3a: Yerevan, Matenaradan, M 196, p. 65a (left) + p. 60d (right); multispectral image, divided,
© Mesrop Mashtots Institute of Ancient Manuscripts (Matenadaran), Yerevan / DeLiCaTe project.
Page 60d contains noteworthy instances of the word ճանապարհ <čanaparh>
(‘road, path’) from Proverbs 2:13 spelled as <čanapah>.

39 Per the Matenadaran assignment, ‘65a’ and ‘60d’ denote additional pages appended to fols 65
and 60; however, the catalogues do not mention them explicitly. See Yeganyan, Zeytunyan and
Antabyan 1965, 270 and Yeganyan, Zeytunyan and Antabyan 1984, 841.

Linguistic Divergence in Armenian Bible and Lectionary Palimpsests  
(3) M 196, p. 60d, ll. 4–5:
{ } [ ]{ } [ ] [ ]
[ ]
Oh {orocՙ} [tՙ]{o}ł[e]al icՙē zčanapah[s owłiłs gnal] i čanapahs [xawa]rins
‘Who leave the paths of uprightness, to walk in the ways of darkness.’
Fig. 3b: Yerevan, Matenadaran, M 196, p. 65a (left) + p. 60d (right), with special spellings marked,
© Mesrop Mashtots Institute of Ancient Manuscripts (Matenadaran), Yerevan / DeLiCaTe project.
Three additional occurrences of the variation between <čanaparh>
and <čanapah> appear on the same page and one more on page 65a,
as well as two instances (one on each of the pages) of <xohowrd>
instead of the expected <xorhowrd> (‘advice, idea, meaning’), which
strongly suggests that the consonant cluster simplification in these cases is not
due to scribal errors (see Fig. 3b).
To ascertain whether the alternation between րհ <rh> and հ <h> is diachronic
or synchronic, recourse to the etymology of the words containing the same conso-
nant cluster could be helpful. An Iranian origin of Armenian ճանապարհ <čana-
parh> was proposed by Henrik Nyberg, followed by Acharyan, who derived it
from Iranian *čarana-parθ-, a compound whose first element is regarded as a
cognate of Old Iranian (more precisely, Avestan: Arm. ‘zenderēn’) kar-, čara-,
čaraya- (‘go’); however, the exact form of the initial compound member, *čarana-,
is not being discussed.40 Acharyan associates the second element with Avestan

40 Nyberg 1928–1931, vol. 2, 187; Acharyan 1926–1935, vol. 3, 183.

  Hasmik Sargsyan
pərətu- (‘hole’) but does not cite any attestations of the compound in Iranian. Addi-
tionally, it remains unclear how the reconstructed *parθ- with the meaning of
‘hole’ might evolve into the meaning of ‘road’ or ‘way’. Birgit Olsen proposes an
alternative Iranian form for the second element (-parh), relying on Harold W.
Bailey; the first element, čana-, remains doubtful for her.41 Ralf-Peter Ritter sug-
gests Armenian čem- as in čemim ‘I strut, I walk’ (< Parthian c’m ‘come’, cm ‘run,
move’) as the basis for the first element of the compound, with a dissimilation of
m to n in the presence of the bilabial p in čanaparh.42
In any case, given the numerous other Iranian loans in Armenian featuring
the same consonant cluster, it is probable that ճանապարհ <čanaparh> also has
an Iranian origin. Giancarlo Bolognesi discusses different dialectal sources for the
Armenian loanwords of Iranian origin with the alternation hr / rh / h.43 He ob-
serves that the Iranian borrowings in Armenian typically exhibit the North-
western Iranian, specifically Parthian, isogloss hr, which in earlier borrowings
experiences either a metathesis hr > rh or a simplification into h. According to
Bolognesi, hr remains unaltered in later Parthian borrowings. Jost Gippert illus-
trates that in Armenian, the outcome of Iranian hr < *θr is rh in loans dating back
to Armenian’s pre-literary times, exemplified by the word աշխարհ <ašxarh>
(‘world, country’).44 Additionally, he notes that Georgian counterparts of Armeni-
an words featuring the Iranian consonant cluster undergo a simplification from
rh > to h in loans of later attestations, particularly of the twelfth century.45
On the other hand, Acharyan documents various modern dialectal forms of
ճանապարհ <čanaparh>, including ճամփա <čampՙa>, ճնապա <čnapa>,
ճանապար <čanapar>, and ճամբախ <čambax>.46 The most prominent variant is
ճամփա <čampՙa>, also spelled ճամբա(յ) <čamba(y)>.47 In this form, possibly
through an intermediate stage ճանապահ <čanapah>, the second a is syncopated

41 Olsen 1999, 892–893; Bailey 1956, 104–107. See also Bailey 1986, 7–8 for a possible Iranian
origin of խորհ <xorh> and խոհ <xoh> (‘thinking, counsel’) and the derivations խորհուրդ <xo-
rhowrd> (‘advice, idea, meaning’) and խորհիմ <xorhim> (‘to think’).
42 Ritter 1986, 310; cf. also his etymological considerations on -parh as a cognate of pa(r)hak
‘guard, watch’ on pp. 308–310, concluding with a meaning of ‘way, path’.
43 Bolognesi 1960, 15–17.
44 Gippert 2005, 148; see also Hübschmann 1895–1897, 14.
45 Alternatively, there is a potential for a third Western Middle Iranian source for rh vs h; see
Korn and Olsen 2012 for more insights.
46 Acharyan 1926–1935, vol. 3, 183; it is also noteworthy that, for example, the colloquial Modern
Eastern Armenian variant, ճանապար <čanapar>, which retains only the ր <r> from this conso-
nant cluster, represents yet another possible outcome.
47 For this spelling, see, for example, Acharyan 1926–1935, vol. 3, 182.

Linguistic Divergence in Armenian Bible and Lectionary Palimpsests  
and the consonant cluster rh is reduced to zero. The passages from the manuscript
M 196 cited earlier could be an attestation of this intermediary stage. Another
attestation of this stage appears to be the dialectal variant ճամբախ <čambax>, in
which the final h has undergone a sound change to x. This could also imply that a
dialectal variation involving the alternation between ճանապահ <čanapah> and
ճանապարհ <čanaparh> cannot be ruled out.
 Conclusions
The study of linguistic variation in the oldest textual witnesses of Armenian will
help better envision the development of the literacy of the earliest stages of Ar-
menian and its contact languages, afford a deeper understanding of language
variation in early and modern times, and contribute to the history of text trans-
mission of Armenian and its contact languages. Systematic analysis of this varia-
tion might reveal, among other things, more about the linguistic varieties of Old
Armenian (fifth to tenth centuries). Reliance on manuscript evidence becomes
essential as we endeavour to discern patterns of orthographic consistency. In this
context, palimpsest fragments offer valuable insights, since they may date back to
earlier times than the oldest dated non-palimpsested manuscripts.
Acknowledgements
This publication is part of a project that has received funding from the European
Research Council (ERC) under the European Union’s Horizon 2020 research and
innovation programme (Grant agreement No. 101019006) and from the Deutsche
Forschungsgemeinschaft (DFG, German Research Foundation) under Germany’s
Excellence Strategy – EXC 2176 ‘Understanding Written Artefacts: Material, Inter-
action and Transmission in Manuscript Cultures’, project no. 390893796. The re-
search was conducted within the scope of the Centre for the Study of Manuscript
Cultures (CSMC) at Universität Hamburg. I wish to thank Jost Gippert and José
Maksimczuk for commenting on an earlier version of this paper, as well as Emilio
Bonfiglio, Caroline Macé, Eka Kvirkvelia and Emmanuel van Elverdinghe for in-
sightful comments.

  Hasmik Sargsyan
References
Abeghyan, Manuk (1936),
Գրաբարի քերականություն
[Grammar of Old Armenian], 2nd edn,
Yerevan: State Publishing House of the Armenian SSR.
Acharyan, Hrachya (1928),
Հայոց գրերը
[The Armenian script], Vienna: Mkhitarist Press.
Acharyan, Hrachya (1926–1935),
Հայերէն արմատական բառարան
[Etymological dictionary of Armeni-
an], 7 vols, Yerevan: Yerevan State University Press. Available online at <http://nayiri.com/imagedDic
tionaryBrowser.jsp?dictionaryId=7&dt=HY_HY&pageNumber=0> (accessed on 9 April 2024).
Aghayan, Eduard (1964),
Գրաբարի քերականություն
[Grammar of Old Armenian], Yerevan: Arme-
nian Academy of Sciences.
Alexanian, Joseph M. (2012), The Ancient Armenian Text of the Acts of the Apostles (Corpus Scriptorum
Christianorum Orientalium, 632; Scriptores armeniaci, 31), Leuven: Peeters.
Amalyan, Hayk (1996),
Գիրք Մակաբայեցւոց. Քննական բնագիր
[The Book of Maccabees: Critical
edition], Yerevan: Armenian Academy of Sciences.
Bailey, Harold W. (1956), ‘Armeno-Indoiranica’, Transactions of the Philological Society, 55/1: 88–126.
Bailey, Harold W. (1986), ‘An Iranian Miscellany in Armenian’, in Dickran Kouymjian (ed.), Armenian
Studies / Études arméniennes: in Memoriam Haïg Berbérian, Lisbon: Calouste Gulbenkian Founda-
tion, 1–8.
Bolognesi, Giancarlo (1960), Le fonti dialettali degli imprestiti iranici in armeno (Pubblicazioni dell’Università
cattolica del Sacro Cuore, ser. 3: Scienze filologiche e letteratura, 1), Milan: Vita e pensiero.
Gippert, Jost (2005), ‘Das Armenische – eine indogermanische Sprache im kaukasischen Areal’, in
Gerhard Meiser and Olav Hackstein (eds), Sprachkontakt und Sprachwandel: Akten der XI. Fachta-
gung der Indogermanischen Gesellschaft, 17.–23. September 2000, Halle an der Saale, Wiesbaden:
Reichert, 139–160.
Gippert, Jost (2010), The Caucasian Albanian Palimpsests of Mount Sinai, vol. 3: The Armenian Layer (Mo-
numenta Palaeographica Medii Aevi, Series Ibero-Caucasica, 2/3), Turnhout: Brepols.
Gippert, Jost (2022), ‘Exploring Armenian Palimpsests with Multispectral and Transmissive Light
Imaging’, Magaghat, <https://www.magaghat.ai/articles/exploring-armenian-palimpsests-with-
multispectral-and-transmissive-light-imaging> (accessed on 16 January 2024).
Gippert, Jost (2023), ‘New Light on the Caucasian Albanian Palimpsests of St. Catherine’s Monastery’,
in New Light on Old Manuscripts: The Sinai Palimpsests and Other Advances in Palimpsest Studies
(Veröffentlichungen zur Byzanzforschung, 45), Vienna: Austrian Academy of Sciences, 187–215
<https://doi.org/10.1553/978OEAW91575s187>.
Gippert, Jost, Wolfgang Schulze, Zaza Aleksidze and Jean-Pierre Mahé (2008), The Caucasian Albanian
Palimpsests of Mount Sinai, 2 vols (Monumenta Palaeographica Medii Aevi, Series Ibero-
Caucasica, 2/1–2), Turnhout: Brepols.
Gyulbudaghyan, Sirak (1973),
Հայերենի ուղղագրության պատմություն
[History of the orthogra-
phy of Armenian], Yerevan: Yerevan State University.
Hübschmann, Heinrich (1876), ‘Ueber Aussprache und Umschreibung des Altarmenischen’, Zeitschrift
der Deutschen Morgenländischen Gesellschaft, 30/1: 53–73.
Hübschmann, Heinrich (1882), Die Umschreibung der iranischen Sprachen und des Armenischen, Leipzig:
Breitkopf und Härtel.
Hübschmann, Heinrich (1895–1897), Armenische Grammatik, vol. 1: Armenische Etymologie, Leipzig:
Breitkopf und Härtel.
Klingenschmitt, Gert (1982), Das altarmenische Verbum, Wiesbaden: Reichert.

Linguistic Divergence in Armenian Bible and Lectionary Palimpsests  
Korn, Agnes and Birgit Anette Olsen (2012), ‘On Armenian -agin: Additional Evidence for a Third West
Middle Iranian Dialect?’, Münchener Studien zur Sprachwissenschaft, 66/2: 201–220, available
online at <https://hal.science/hal-01340802/document> (accessed on 19 January 2024).
Kouymjyan, Dickran (2015), ‘Armenian Manuscripts’, in Alessandro Bausi, Pier Giorgio Borbone,
Françoise Briquel-Chatonnet, Paola Buzi, Jost Gippert, Caroline Macé, Marilena Maniaci, Zisis
Melissakis, Laura E. Parodi and Witold Witakowski (eds), Comparative Oriental Manuscript Studies
(COMSt): An Introduction, Hamburg: Tredition, 38–40.
Künzle, Beda O. (1984), Das altarmenische Evangelium / L’Evangile arménien ancien, 2 vols (Europäische
Hochschulschriften, ser. 21: Linguistik und Indogermanistik, 33), Bern: Peter Lang.
Meillet, Antoine (1903), ‘Observations sur la graphie de quelques anciens manuscrits de l’évangile
arménien’, Journal asiatique, 10e s., 2/[3]: 487–507.
Nyberg, Henrik S. (1928–1931), Hilfsbuch des Pehlevi, vol. 1: Texte und Index der Pehlevi-Wörter; vol. 2:
Glossar, Uppsala: Almqvist & Wiksell.
Olsen, Birgit A. (1999), The Noun in Biblical Armenian: Origin and Word Formation, With Special Emphasis on the
Indo-European Heritage (Trends in Linguistics: Studies and Monographs, 119), Berlin: De Gruyter.
Renoux, Athanase (1971), Le codex arménien Jérusalem 121, vol. 2: Édition comparée du texte et de deux
autres manuscrits / introduction, textes, traduction et notes (Patrologia Orientalis, 36/2), Turnhout:
Brepols.
Ritter, Ralf-Peter (1986), ‘Armenisch pahak ownim, taraparhak varem “ἀγγαρεύω”’, Die Sprache, 32:
308–310.
Schmitt, Rüdiger (1972), ‘Empfehlungen zur Transliteration der armenischen Schrift’, Zeitschrift für
vergleichende Sprachforschung, 86/2: 296–306.
Stone, Michael E. (1993), ‘Variants in Armenian Manuscripts and Their Assessment’, in Henning Leh-
mann and Jos J. S. Weitenberg (eds), Armenian Texts: Tasks and Tools (Acta Jutlandica, 69/1; Hu-
manities Series, 68), Aarhus: Aarhus University Press, 15–25.
Weitenberg, Jos J. S. (1994), ‘Text Chronological Aspects of the Use of the Article -ս and -դ in Gospel
Manuscripts M and E’, in Shahe Ajamian and Michael E. Stone (eds), Text and Context: Studies in
the Armenian New Testament, Papers Presented to the Conference on the Armenian New Testament,
May 22–28, 1992 (University of Pennsylvania Armenian Texts and Studies, 13), Atlanta: Scholars
Press, 97–113.
Weitenberg, Jos J. S. (1996), ‘On the Early Development of the Armenian Dialects’, in Dora Sakayan
(ed.), Proceedings of the Fifth International Conference on Armenian Linguistics, McGill University
Montreal, Quebec, Canada, May 1–5, 1995, Delmar, NY: Caravan Books, 93–118.
Weitenberg, Jos J. S. (1997), ‘The Prepositional Group
ի
յ
- i y- and the Orthography of Gospel Manu-
script M (Matenadaran 6200)’, Annual of Armenian Linguistics, 18: 39–50.
Weitenberg, Jos J. S. (2006), ‘Aspects of Classical Armenian Orthography: Armenian e, ē and the Greek
Names in the Gospels’, in Anna Krasnowolska, Kinga Maciuszak and Barbara Mękarska (eds), In
the Orient where the gracious light … Satura Orientalis in honorem Andrzej Pisowicz, Kraków:
Księgarnia Akademicka, 215–228.
Weitenberg, Jos J. S. (2014), ‘Manuscripts and Dialects’, in Valentina Calzolari and Michael E. Stone (eds),
Armenian Philology in the Modern Era: From Manuscript to Digital Text (Handbook of Oriental Stud-
ies, sect. 8, 23/1), Leiden: Brill, 214–225.
Yeganyan, Onnik, Andranik Zeytunyan and Paylak Antabyan (1965),
Ցուցակ ձեռագրաց Մաշտոցի
անվան Մատենադարանի
[Manuscript catalogue of the Mashtots Instute of Ancient Manu-
scripts (Matenadaran)], vol. 1, Yerevan: Armenian Academy of Sciences, available online at
<https://www.matenadaran.am/ftp/Hamarot/Dzeragrats-1.pdf> (accessed on 16 January 2024).

  Hasmik Sargsyan
Yeganyan, Onnik, Andranik Zeytunyan and Paylak Antabyan (1984),
Մայր ցուցակ հայերէն
ձեռագրաց Մաշտոցի անուան Մատենադարանի
[Catalogue of Armenian manuscripts of
the Mashtots Institute of Ancient Manuscripts (Matenadaran)], vol. 1, Yerevan: Armenian Acad-
emy of Sciences, available online at <http://www.matenadaran.am/ftp/data/M/Hator-1.pdf> (ac-
cessed on 9 May 2024).
Zōhrapean, Yovhannēs <Zohrab> (1805),
Աստուածաշունչ
․
Մատեան Հին եւ Նոր Կտակարանաց
[The Bible: Old and New Testaments], 4 vols, Venice: San Lazzaro.