PHONETIC CHANGE IN MEDIEVAL GREEK: FOCUS ON LIQUID INTERCHANGE *
anolessou & Notis Toufexis University of Patras, University of
Cambridge [email protected], [email protected]
Abstract The present paper aims to propose a methodology for the
research of phonetic change in Medieval Greek, and to show, through
the investigation of the phenomenon of liquid interchange, that a)
the new electronic tools available for the study of Greek can
contribute crucially to the research on the distribution of
phonetic changes and b) phonetic changes in Medieval Greek are to a
large extent regular and conform to cross-linguistic patterns.
1. Introduction Medieval Greek presents no major regular
changes, i.e. conditioned changes of the Neogrammarian type, which
apply without exceptions (or with explicable exceptions) in all the
potential phonetic environments. Thus, whereas the student of the
history of Ancient Greek meets with a standard list of regular
phonetic changes with precisely describable conditions, which led
from Indo-European to Greek or from Proto-Greek to the various
Ancient Greek dialects (e.g. /a:/ > /:/ in Ionic and Attic) 1
and even from Attic Greek to the Koine, the phonetic rules of
Medieval Greek never seem to apply across the board, resulting in
considerable variation both in the texts of the period (Manolessou
2008) and Modern Greek (and its dialects). The reaction of scholars
to this issue has up to now been one of hopeful procrastination: to
postpone discussion pending a better knowledge of a) Medieval Greek
texts and b) Modern Greek dialects. Now it is true that studies of
MG dialectology are lagging behind in comparison with those in
other countries. However, this is mainly true of theoretical/
structural descriptions of dialects (their phonology, morphology,
syntax etc.) 2 . It is not true of dialectal vocabularies, as there
exist hundreds of (amateur or not) local glossaries, and a wealth
of dialectal attestations for each lexeme of Modern Greek collected
in the archives of the Academy of Athens, dated
*
This paper is the outcome of research conducted for the project
Grammar of Medieval Greek at the University of Cambridge, funded by
a grant from the Arts & Humanities Research Council. For more
information see
http://www.mml.cam.ac.uk/greek/grammarofmedievalgreek . 1 These
changes are described in standard historical grammars of Ancient
Greek, e.g. Rix (1992). 2 For a recent overview of theoretical
approaches to Modern Greek dialects, see Ralli (2006).
291
and geographically localised 3 . The situation nowadays,
therefore, is quite different from the time when the first
descriptions of Medieval Greek phonology were written (on which the
descriptions of modern handbooks are based cf. the overview of
desiderata in Mirambel 1953): it is now possible to track the
spread of any phonetic change lexeme by lexeme in the Modern Greek
dialects. Roughly the same observation can be made concerning our
knowledge of Medieval Greek texts: a very large amount of published
texts, both literary and non-literary, not available to earlier
scholars is now easily accessible. We also possess better
linguistic descriptions of Medieval Greek in general or of specific
texts in particular, better tools for textual searches (such as
electronic searchable texts, concordances, databases etc.) 4 , all
of which were not at hand for those who first tackled the vagaries
of Medieval Greek phonology. It is now possible therefore, as it
was not before, to examine the spread and distribution of each
phonetic change in all major and hundreds of minor medieval texts,
from different areas and different periods 5 . Nevertheless,
despite the considerable improvement in the available data, the
current linguistic scholarship on Greek contains little discussion
of specific phonetic changes in Medieval Greek. One reason for this
apparent lack might be the fact that the better knowledge of
Medieval Greek texts and MG dialects has not brought about the
better understanding of the precise (phonetic)
conditions/environments of application of a phonetic change (for
example, why and but not *; why and but not *; 6 ) that researchers
had hoped for. Thus, thorough examination of large amounts of
medieval material by the Cambridge Grammar of Medieval Greek has
shown that: a) no matter how vernacular a text, from whatever
period or area, medieval sound changes are almost never attested
with complete normality, i.e. without exceptions. For example, Dig.
G III.69 but Dig. G IV.465 ; Chron.Mor. H 6646 but Chron.Mor. H
633; chill.O 279 but chill.O 551; Liv.V 2443 but Liv.V 1429. There
is always variation, sometimes more, sometimes less. b) most
inherited words which in Standard Modern Greek exist in unchanged
form may appear in MedG with the change. For example, the following
are all MedG attestations of the
On the contents of the Archives of the Center for Modern Greek
Dialects of the Academy of Athens, see Manolessou, Karantzi &
Yakoumaki (2004), and on its electronic availability see Academy of
Athens (2001) and the website www.academyofathens.gr/ilne (accessed
12/03/2008). An electronic bibliography of local dialectal
glossaries is available from the , at
http://www.greek-language.gr/greekLang/medieval_greek/bibliographies/idiomatic/contents.html
(accessed 12/03/2008). 4 On electronic corpora in historical Greek
dialectology, see Manolessou & Toufexis (forthcoming), on
available electronic resources in the field of Medieval Greek see
Toufexis (forthcoming). 5 With all the caveats, of course,
concerning the reliability of written sources of past periods, and
especially Medieval Greek, for the representation of authentic
linguistic usage (Manolessou 2008). 6 On this specific change (/i/
> /u/ adjacent to labial and velar consonants and s) see Joseph
(1979), Moysiadis (2005: 97-103-105) and references therein.
3
292
change e > i adjacent to liquids and nasals, which do not
appear so in MG: (1473, Corfu, KONIDARI & RODOLAKIS 1996: 78,
186.3); (1073, Miletos, VRANOUSI 1980: 50, 9.105); (1148, Calabria,
GUILLOU 1963: 7, 82.17); (1112, Sicily, CUSA 1868/82: XIV, 410.17);
Poulol. 540 app. cr. (); (1581, Andros, POLEMIS 1995: 26, 165.19);
MACH., Chron. 81.12 (Pieris/Nikolaou-Konnari); Chron. Toc. 703;
Diig. Alex. E 233.3 (Lolos). c) almost any inherited word which in
Modern Greek has undergone the change may appear in Medieval Greek
without it, even in very vernacular texts: For example, the
following are attestations of forms which in MG have undergone
synizesis (which dates probably around the 13th c. (Minas 1983:
289)), but which appear in Early Modern Greek texts in unchanged
form, although the word itself (and the text from which it is
taken) is vernacular: (1635, Nisyros, TSIRPANLIS 1982: 1, 13.3);
BERG., Apok. V 27; NOUK., Aisop. Myth. 28.2; (1500, Crete,
MAVROMATIS 1994: 1 [], 237.15); , , , , , (1642, Patmos,
MICHAILARIS 1998: 1, 193.12). d) there is a vast amount of
hypercorrection going on, which means i) that frequently the change
is attested only indirectly, through changes in etymologically
unjustifiable words, something which makes both electronic and
traditional searches for attestations of the spread of the
phenomenon very difficult and ii) that the tendency to correct and
hide the change in written language is very strong. For example,
the following are attestations of hypercorrection of the Northern
dialect phenomenon of mid-vowel raising: (1445, Macedonia, BOMPAIRE
1964: 30, 217.30); (post 1356?, Epirus, ALEXOULIS 1892: 1, 277.7);
(1618, Serres, ODORICO 1998: 33, 114.3); (1695, Kastoria, MERTZIOS
1947: 2, 212.4). The cause of the variation and of the
non-regularity of MedG phonetic change is not that we do not
understand the environments, the conditions, and the working of the
changes, or that we do not have more precisely geographically and
chronologically located information on the spread of the changes.
Also, it is not that they are by nature sporadic changes (like
metathesis or vowel assimilation). On the contrary, many of them
are perfectly regular neogrammarian sound changes, which
predictable environments and exceptions (analogical levellings,
hypercorrections, posterior loans). Its just that they were
arrested in progress. Medieval (and Modern) Greek is an ideal
exemplification and documentation of spread of sound change by
lexical diffusion. In some areas, in some periods, in some
registers, more words are affected than in others. Ultimately, the
period of activity of the change is over, before it has covered the
totality of the vocabulary. The parallel existence of a learned
language as a model could be seen as the ultimate inhibiting
factor. Frequently it can be shown that the period of activity of
a
293
phonetic change is several hundred years, or that in fact it is
still active in Modern Greek, and still being counteracted by the
standard language. This of course does not entail that it is
fruitless to investigate MedG source texts from the point of view
of phonology. First of all, we do need the data in order to rule
out the possibility that the conditioning factor is indeed a
specific phonetic environment. Secondly, we need up to date
descriptions of changes, couched in modern terms, so as to be
comparable with the results of research on other languages and to
contribute to linguistic theory on phonetic change. And thirdly, we
need to have an accurate picture of the various phonetic changes in
Medieval Greek not only for use in comparative historical
linguistics, but also in sociolinguistics, and textual edition and
interpretation. 2. A case study: liquid interchange In order to
exemplify concretely the claims made above concerning a) the
feasibility and utility of research on and b) the interpretation of
the regularity (or non-regularity) of phonetic change in Medieval
Greek, in what follows, we will investigate a specific sound
change, the so-called liquid interchange, which in fact consists of
two different sub-changes: a) b) change of /l/ > /r/ when it is
the first member of a consonant cluster (classic examples: > ,
> ) and change of /r/ > /l/ when another /r/ sound follows in
the same word (classic examples: > , > ). We aim to provide
as much detail as possible, i.e. by examining all lexemes of the
Greek language possessing the appropriate phonetic environments,
and checking from both written and oral sources (medieval documents
and dialectal data) whether they are attested in changed form or
not. The first to have noted and described the phenomenon of liquid
interchange is Simon Portius in his grammar of 1638 (Meyer 1889:
10). The phenomenon is of course well-known in the literature on
the history of Greek and its dialects, but there are in fact only
one and-ahalf article dedicated to it: a detailed setting out of
the data by J. Psichari (1905/1930), collecting most known (at the
time) instances of the phenomenon from Ancient, Medieval and Modern
Greek, and a short attempt at interpretation by G. Shipp (1958),
linking it with other dissimilatory phenomena in Greek phonology.
We are still lacking, therefore, a modern account which will
incorporate a) cross-linguistic data and theoretical phonological
accounts and b) the new wealth of information available through the
progress of research on Greek and its dialects, as well as the new
electronic research tools.
294
2.1. Cross-linguistic and theoretical documentation Concerning
the first point, both sub-phenomena of liquid interchange, i.e.
dissimilation of consecutive liquids and liquid variation in front
of consonants, are well-attested in many languages. The behaviour
of Greek, therefore, has to be compared with similar
crosslinguistic attestations and the interpretations proposed for
them. In detail: The most well-known case of liquid dissimilation
is that of Latin, in which the derivational suffix -/alis/ appears
as -/aris/ when the stem of the word contains another /l/ (Steriade
1987, Odden 1994 among others): navis nav-alis, natura natur-alis,
sol sol-aris, linea line-aris A similar case of dissimilation in
specific suffixes occurs in Georgian (Fallon 1993), where the
derivational suffix -/uri/, forming ethnic adjectives, appears as
-/uli/ when the stem contains another /r/: dan-uri Danish,
polon-uri Polish, ungr-uli Hungarian, aprik-uli African Yet a third
documented case is the West Indonesian language of the Sunda
islands (Sundanese), in which the plural infix -/ar/ appears as
-/al/ when the root contains another /r/ (Cohn 1992, Holton 1995):
kusut k-ar-usut messy, dama d-ar-ama well, hormat h-al-ormat
respect, combrek c-al-ombrek cold All of these phenomena are in
fact grammaticalised dissimilations, in that they require
grammatical, i.e. morphological information, since they concern
specific affixes. The Greek case is somewhat different: it affects
consecutive /r/ sounds in the root of words, and does not seem to
occur in suffixes or in composition (see below for a more detailed
description of the data). A greater similarity to the Greek
phenomenon is observable in the sporadic dissimilation of
consecutive /r/ sounds in several Romance languages such as
Spanish, Catalan, Portuguese and Italian (Lloret 1997, Colantoni
& Steele 2005): Lat. cerebrum > Sp. celebro brain, Lat.
peregrinum > It. pelegrino pilgrim, Lat. armarium > Port.
almario cupboard, Lat. aratrum > W. Cat. aladre plough. A few
Romance dissimilated words have also made it into English as
loanwords, such as marble ( veldad, Lat. carta > Sp. carta >
calta, Lat. populum > pueblo > puebro, Bermudez (pr. name)
> Belmudez, Lat. palma > Sard. parma > prama, Lat. plus
> Sard. prus, Lat. flamma > Sard. framma. The diachrony of
the Romance languages shows a frequent phenomenon of liquid
interchange in consonant clusters, which may or may not have
survived in the modern standard language. From a theoretical
phonological point of view this phenomenon is interpreted as a
neutralization of the distinctive feature which distinguishes
rhotics from laterals in specific positions (the phenomenon is
technically termed positional neutralization of liquids, Rice
2005). The choice of the surfacing liquid (/r/ or /l/) seems to be
languagespecific, and connected with the notion of markedness: in
each phonological system, one of the liquids is the unmarked one
(the most frequent, with the wider phonotactic distribution) and
this is the one that is preferred in consonant cluster alternations
a result known as emergence of the unmarked in theoretical
phonology. This does not exclude the possibility of the reverse
change also occurring in the language, but in this case it should
perhaps be interpreted as a hypercorrection. In the Greek case, the
unmarked consonant seems to be /r/, as there are far more clusters
with /r/ + consonant than with /l/ + consonant. This can be
(imperfectly) demonstrated by a search in the electronic
296
version of the of the Triantafyllidis foundation 7 , through
comparing the number of words containing the two types of consonant
clusters:Table 1. Frequency of liquid consonant clusters in
Standard Modern Greek v z k l m n ks p r s t f x ps Total /r/ +
consonant 114 631 264 9 171 272 84 620 363 18 140 0 122 681 216 448
4 4157 /l/ + consonant 36 47 7 4 10 119 0 128 25 3 70 0 20 182 69 6
2 728
It seems again, therefore, that Greek can be brought into line
with several other languages presenting similar phenomena, and that
such comparative research can shed light on the causality of the
change. It furthermore seems that liquid alternations in consonant
clusters also possess the potential to become a regular,
exceptionless sound change. 2.2 Historical overview Let us now turn
to the detailed examination of the two liquid interchange phenomena
in the history of Greek. 2.2.1 Classical and Koine Greek here are
no attestations of the consonant cluster interchange in Classical
literature. The first, and quite rare, attestations of the
phenomenon appear in Attic inscriptions, all of which are quite
late, dated to the Roman period. Examples are collected in
Dieterich (1898: 107), Jannaris (1898: 187), Meisterhans-Schwyzer
(19003: 83), Lademann (1915: 119), Schwyzer (1939: 213) and
Psichari (1905/1930). Some of these, however, are misreadings of
earlier scholars or refer to outdated editions; the most reliable
information comes from Threatte (1994: 483):
See
http://www.greek-language.gr/greekLang/moderngreek/tools/lexica/triantafyllides/index.html
(accessed 12/03/2008). We are fully aware of the methodological
constraints of using only one electronically available lexical
database; the data presented here can without doubt be enriched
with the help of other electronic or traditional resources.
7
297
IG II22242.33, 238/9 AD (< Calpurnius), IG II2 2245.222,
262/3 AD, IG II2 13224.6, 3rd c. AD (< ), gora inv. no. IL
493.27, 3rd c. AD, G III 3531 (early Christian).Interchange in
consonant clusters occurs in the papyri of the Roman and Byzantine
periods (examples from Gignac 1976: 105): PRyl. 160c, i.7,16 (AD
32), PMich.204.5 (AD 7), POxy 1059 (6th AD). However, liquid
interchange in the late Egyptian papyri is more frequent than in
any other written documents of any period of Greek, and furthermore
it happens also unconditionally, i.e. even outside consonant
clusters: , (Pryl 160c, i.10 (AD 32), PMich.310.16 (27 AD), SB 5110
(AD 42). Gignac (1976: 106-7) claims that this frequent interchange
of and indicates that there was only one liquid phoneme /l/ in the
speech of many writers in the Roman and Byzantine periods. In the
literature of this era we also find one or two examples of the
dissimilatory change, namely the frequently occurring (< Lat.
flagellum) in the NT (Ev. Jh 2.15, Matth. 27.26) and nce (Act.
Apost. 27.30 (Sin.)). Summing up this period: both changes begin to
appear around the 2nd 3rd c. AD, with isolated exceptions such as
which do not constitute prototypical instantiations of the
phenomena. It cannot be claimed that the changes possess any
regularity, with the possible exception of Egyptian Greek, where
there is substratum interference. 2.2.2 Medieval Greek oth
phenomena are attested in Medieval Greek, starting from the
sporadic attestations of the early Byzantine period collected by
Psaltes (1913: 76, 98), e.g. DAI 75.11. /l/ > /r/ in consonant
clusters An electronic search of the corpus of Christian
Inscriptions (PHI#6 CD-ROM) has shown that a few specific words
appear frequently in changed form, therefore the change seems to
have been stabilized, for these words at least. For example, the
word and its derivatives appears 16 times, the word and its
derivatives appears 14 times: TAM II 1142; MAMA 4 195; IK Iznik
576; Corinth 8,1 207; SEG 37:1292; IScM III 194; MAMA 3 514.
298
A search of all /l/ + consonant combinations in the electronic
version of Kriaras (1967-) (only up to the lemma ) 8 , plus all
data collected by the Medieval Greek Grammar project has given a
small corpus of 290 words which potentially could undergo the
change. Of these 76 are attested in Medieval Greek as actually
having undergone it (percentage 26 %): , , , , , , , , , , , , , ,
, , , , , , , , , , , /, /, , , , , , , , , , , , , , , , , , , ,
/, , , , , , , , , , , , , , , , , , , , , , , , , , , , , . An
important note that should be made is that the /l/ > /r/ change
is an active change in Medieval Greek, as it has affected not only
forms inherited from previous periods, but also medieval loanwords
(such as , , , , ). The examination of the words that have not
undergone the change shows that they are mostly learned words, such
as: , , , , , , , , , , , , , , , , , , , , , . Also some of them
are hapax words in the medieval literature, such as: , , ,
.
8
Available under
http://www.greek-language.gr/greekLang/medieval_greek/kriaras/index.html
(accessed 12/03/2007).
299
There are a number of words which seem to have undergone the
reverse change, i.e. changing a /r/ + consonant cluster to a /l/ +
consonant cluster: , , , , , , , , , , , , B, , , , , , . These
constitute less than 1% of the total words which could potentially
undergo the change, i.e. the medieval words which contain a cluster
of /r/ + consonant. Furthermore, for some of them alternative
causes can be brought forward: for example, the /l/ in (< it.
arbitro < lat. arbiter) and its family of words could be
attributed to dissimilation with the following second /r/ in the
word, or could have been borrowed already containing the /lb/
cluster, as this form is attested in Italian dialects such as
Venetian. In this connection it should be noted that most of these
words are of foreign origin ( < burrichus, < burgensis, <
purgatorium, < veretone). Note also the proper name Bo in the
Alexiad < Burkhard (3.10.4.13, ed. Leib). Interestingly, Portius
already had noted that it is foreign words which tend to change /r/
to /l/ in consonant clusters, giving the example of < scrima
(Meyer 1889: 10). On the basis of the above, therefore, it can be
claimed that the reverse /r/ > /l/ change is not a regular
phenomenon of Greek, but a hypercorrection or an imitation of
Romance models. he medieval data can be complemented through the
evidence of MG dialects, from which a further 22 words presenting
the change /l/ > /r/ can be added to the above, bringing the
total to 98 (percentage 34 %): , , , , , , , , , , , , , , , , , ,
, , , . Dissimilation Turning now to the second phenomenon, that of
the dissimilation of consecutive rhotic sounds, even a cursory
examination shows that it is much less widespread than the first.
Thus, of a total of 831 words (in Kriaras Lex. and the data
collected by the Medieval Grammar project) that could potentially
undergo the change, only 30 of them do, again either in MedG source
texts or in MG dialects (percentage 3,61 %):
300
, , , , , , , , , , , , , , , , , , , , , , , , , /l/ or /l/ to
/r/ change in that the liquid constitutes the second member of the
cluster, or there is no second liquid in the word. Examples include
, . As far as the distribution in Medieval source texts is
concerned, it can be stated that there are no sources where the
change appears with complete regularity, meaning both that a) the
same lexical item may appear in the same text in both the changed
and unchanged form and b) different lexical items, both presenting
the proper phonological environment for the change behave
differently in the same document, one presenting the change and one
not. Interestingly, the same picture is presented by Spanish
documents of the same period, exhibiting the same phenomenon of
liquid interchange (Fontanella de Weinberg 1984). To give one
characteristic example, the notary and professional scribe uses in
the 154 wills (dated 1506-1532) that are available to us today (ed.
Mavromatis / Lambakis 2003) always only the unaltered from of his
surname (-); the altered form - appears in these documents only
twice: once in the signature of a witness (no. 2 , p. 6) and once
in a written instruction of the duke of Candia to himself ( | ...,
p. 65). It is obvious that while the sound change was operating in
Crete at the time this notary was writing, he chose not to
represent it in the written records he is producing. The key factor
which seems to increase or decrease the percentage of changed items
is register. Thus, again, documents from Cyprus, which are in a
more decidedly dialectal character, present both greater regularity
in same words, and a wider variety of words which show the change
some of the words in the lists provided above are attested only in
Cypriot documents (e.g. , ).
301
2.2.3 Modern Greek An electronic search of the /l/ + consonant
clusters in the Triantaphyllidis dictionary has given a total of
460 words which could potentially undergo the change. Of these,
only 19 words appear in changed form in Standard Modern Greek: , ,
, , , , , , , , , , , /, , , , , . However, if one turns to the MG
dialects, and more specifically to the dialectal archives of the
Academy of Athens, an impressive additional 130 words are found,
bringing the percentage of SMG words presenting the phenomenon from
to 32%, similar to the levels of Medieval Greek: , , , , , , , , ,
, , , , , , , , , , , /, ( , > , > ) which would make the
total considerably higher. The words that have not undergone the
change are almost always either learned new creations or recent
loanwords:
302
, , , , , , , , , , , , , etc. On the basis of the above it is
possible to conclude that the change is regular in MG, as well,
leaving unaffected only the words belonging to higher registers. It
must be noted that the change is still active in the dialects,
because even comparatively recent loanwords have undergone it,
including: revolver > > (Corfu), (Eub., Chios), (Cephall.),
(Pont.), (Peloponn.), (Ionia), waltz > > (Naxos), bolshevik
> > (Karpathos). Although the sampling of the dialectal
material available in the Academy of Athens is unequal (i.e. some
areas are overrepresented), it is apparent from the distribution of
the changed forms that the phenomenon is stronger in southern
dialects, such as Cycladic, Cretan and Dodecanesian. As far as the
second change is concerned, that of dissimilation of /r/ to /l/ in
the presence of a second rhotic, the data are again similar to the
medieval ones. That is, the number of words that could potentially
undergo the change, as they contain two /r/s, is very high, about
3400. However, in standard MG less than 10 words appear in changed
form, a percentage of less than 0,2 %: , , (marginal), , , , The MG
dialects add to this meager number an additional 20 words, which
might increase if we investigate dialectal material more
thoroughly: (< erba rosa, ), , , , , , , , , (< ), (< ),
(< ), < , < renper, < , < , , , , . Again, the
presence of comparatively recent lexical items which have undergone
the change (e.g. ) shows that the change is still active in MG. As
far as one can tell from the examination of the items which have
not undergone the change, the limitation to non-learned high
register vocabulary is not sufficient here. The change is not
apparent also in the following contexts:
303
a) when the word is a compound. For example, none of the
hundreds of compounds in -, -, -, - displays the change. he Greek
phenomenon operates in a different way from the grammaticalised
suffix interchanges seen in other languages. b) when the change
would create a consonant cluster disallowed or dispreferred by the
phonotactics of Greek, e.g. > *, > ? 3. Conclusions We have
investigated two different sound changes in the diachrony of Greek.
Both of them involve liquids, the first in consonant clusters and
the second in dissimilatory contexts. It has been shown that the
detailed examination of the data, with the help of electronic
research tools, can help us draw clearer conclusions as to the
behaviour of sound changes in Greek: both changes are active for a
very extensive period of time (about 2000 years), something that
requires explanation. Also, they are both cross-linguistically
well-attested changes which can be explained on the basis of known
phonological processes. However, in their distribution they are
widely dissimilar: while the first change is a potentially regular
change, spreading to about one third of the vocabulary of any given
period, and arrested primarily in lexical items of higher
registers, the second change is a sporadic one, affecting less than
1% of the vocabulary. We hope to have shown that a) the specific
change investigated here can provide a useful case study for
research on the diffusion of sound change in general and b) the
methodology followed for the present research can, and should, be
applied to all other sound changes affecting Medieval and Modern
Greek. ReferencesAcademy of Athens, (1933-). , . Athens. cademy of
Athens, (2001). 23 - . Achil. O = Smith O. L., (1990). The Oxford
Version of the ACHILLEID [Opuscula graecolatina, 32], Copenhagen.
Alexoulis, A., (1892). , 4 (1892) 275-281. Apok. V = Vejleskov P.,
(2005). APOKOPOS. A fifteenth century Greek (Veneto-Cretan)
catabasis in the vernacular. Synoptic edition with an introduction,
commentary and Index verborum, Cologne: Romiosini. Bompaire J.,
(1964). Actes de Xropotamou. dition diplomatique, Paris. Chron.
Mor. = Schmitt J., (1904). The Chronicle of Morea, . A History in
Political Verse [...], London 1904, repr. Groningen 1967, Athens
2003 Chron. Toc. = Schir G., (1975). Cronaca dei Tocco di Cefalonia
di anonimo: prolegomeni, testo critico e traduzione, a cura di G.
Schir, Rome 1975. Cohn A., (1992). The Consequences of
Dissimilation in Sundanese. Phonology 9, 199-220. Colantoni L.
& J. Steele (2005). Liquid asymmetries in French and Spanish.
Toronto Working Papers in Linguistics 24, 1-14. Cusa S., (1868/82).
I diplomi greci ed arabi di Sicilia, Palermo vol. 1. part 1, 1868,
vol. 1. part 2, 1882, repr. Cologne Vienna: Bhlau 1982. Dieterich
K., (1898). Untersuchungen zur Geschichte der griechischen Sprache
von den hellenistischen Zeit bis zum 10. Jahrh. n.Chr. Leipzig:
Teubner. Dig. G = Jeffreys . ., (1998). Digenis Akritis. The
Grottaferrata and Escorial Versions, Cambridge: Cambridge
University Press 1998.
304
Diig. Alex. E = Lolos A. C., (1983). Ps.-Kallisthenes: Zwei
mittelgriechische Prosa-Fassungen des Alexanderromans, Teil I,
Knigstein 1983. Fallon P., (1993). Liquid Dissimilation in
Georgian. Proceedings of the Tenth Eastern States Conference on
Linguistics (ESCOL 93), ed. by Andreas Kathol and Michael
Bernstein, 105-116. Fontanella de Weinberg, M. B., (1984). Confusin
de lquidas en el espaol rioplatense (siglos XVI a XVIII). Romance
Philology 37, 432-45. Frigeni, Ch., (2005). The development of
liquids from Latin to Campidanian Sardinian: The role of contrast
and structural similarity. Toronto Working Papers in Linguistics
24, 1530 Gignac, F. T., (1976). A Grammar of the Greek papyri of
the Roman and Byzantine Periods. Vol. I, Phonology. Milano:
Cisalpino-Goliardica. Guillou A., (1963). Les actes grecs de S.
Maria di Messina. Enqute sur les actes grecs dItalie du sud et de
Sicile (XI e-XIV e s.), Palermo 1963. Holton, D., (1995).
Assimilation and Dissimilation of Sundanese Liquids. In J. Beckman,
L. Dickey and S. Urbanczyk (eds.), Papers in Optimality Theory:
University of Massachusetts Occasional Papers 18, 167- 180.
Jannaris, A.N., (1897). An historical Greek Grammar chiefly of the
Attic dialect as written and spoken from classical antiquity down
to the present time, founded upon the ancient texts, inscriptions,
papyri and present popular Greek. London/ NY: MacMillan. Joseph B.
D., (1979): Irregular [u] in Greek. Die Sprache 25, 46-8. Karaboula
D.P. & L. Paparriga-Artemiadi, (1998). (1513-1516), 34 (1998)
9-126. Konidari I. M. & G. Rodolakis, (1996). (1472-1473), 32
(1996), 139-205. Lademann, W., (1915). De titulis atticis
questiones orthographicae et grammaticae. Kirchhain: Schmersow.
Lemerle P., (1945/46). Actes de Kutlumus. d. diplomatique, Paris,
1945-46. Liv. V = Lendari T., (2007). (Livistros and Rodamne). The
Vatican Version. Critical edition with Introduction, Commentary and
Index-Glossary. Editio princeps, Athens: MIET 2007. Lloret, M. R.,
(1997). Sonorant dissimilation in the Iberian languages. In F.
Martnez-Gil & A. Morales Front (eds.), Issues in the phonology
and morphology of the major Iberian languages, Washington D. C.:
Georgetown University Press, 127-50. Mach., Chron. O = Pieris M.,
Nikolaou-Konnari A., (2003). , . , Nicosia 2003. Manolessou, I.,
(2008). On historical linguistics, linguistic variation and
Medieval Greek. Byzantine and Modern Greek Studies 32, 63-79.
Manolessou I., Ch. Karantzi & E. Yakoumaki, (2004). ( ) . In
Proceedings of the 6th International Conference of Greek
Linguistics (ICGL 6) Rethymnon [available online, accessed
12/03/2008:
http://www.philology.uoc.gr/conferences/6thICGL/ebook/a%5Cmanolessou&giakoumaki&karantzi
.pdf] Manolessou, I. & N. Toufexis, (forthcoming). Corpus
linguistics and historical dialectology. A case study of Cypriot.
In Proceedings of the 3rd International Conference on Modern Greek
dialects and Linguistic Theory, Nicosia, 14-16 June 2007.
Mavromatis, G., (1994). , . (1496-1543), Venice 1994. Mavromatis,
G. & St. Lambakis, (2003). , 1506-1533: , , , Herakleion.
Meisterhans, K. & E. Schwyzer, (19003). Grammatik der attischen
Inschriften. Berlin: Weidmann. Mertzios K. D., (1947). (1695-1699),
[ , 7], Thessaloniki 1947, 209-264 Meyer, W., (ed.) (1889). Simon
Portius, Grammatica Linguae Grecae Vulgaris, 1638. Paris: Vieweg.
Michailaris, P., (1976). (1695-1696) . . . . , 13 (1976), 245-257.
Minas, K., (1983). e/, i/ . 12 (1983) 283-9. Mirambel, A., (1953).
Essai de phonologie du grec byzantin. In Atti del VII congresso
internazionale di studi bizantini 1951 (Studi bizantini e
neoellenici), v.1, 303. Moysiadis, Th., (2005). . . Athens:
Ellinika Grammata.
305
Nouk., Aisop. Myth. = Parasoglou, G. M., (1993) . . . . Athens:
Ermis 1993. Odden, D., (1994). Adjacency Parameters in Phonology,
Language 70, 289-330. Odorico, P., (1998). Le codex B du Monastre
Saint-Jean-Prodrome (Serrs). B (XVe-XIXe sicles), Paris 1998. PHI#6
= Packard Humanities Institute CD-ROM 6. Also available online at
http://epigraphy.packhum.org/inscriptions/ (accessed 12/03/2008).
Polemis, D. I., (1995). . 16 ., Andros. Psaltes, S., (1913).
Grammatik der byzantinischen Chroniken. Gttingen: Vandenhoeck &
Rupprecht. Psichari, J., (1905/1930). Essais de grammaire
historique: Le changement de l en r devant consonnes en grec
ancien, mdieval et moderne. In Psichari J., Quelques travaux de
linguistique, de philologie et de littrature hellniques, 1884-1928.
Paris : Belles Lettres, 664-710. [= Memoires Orientaux, Congrs de
1905, Paris 1905, 231-36]. Poulol. = Tsavari, ., (1987). O . K , ,
thens 1987. Ralli, A., (2006). Syntactic and morphosyntactic
phenomena in Modern Greek dialects. The state of the art. Journal
of Greek Linguistics 7, 121-159. Rice, K., (2005). Liquid
relationships. Toronto Working Papers in Linguistics 24, 3144. Rix,
H., (19922). Historische Grammatik des Griechischen. Darmstadt:
Wissenschaftliche Buchgesellschaft. Sadler, J. D., (1973).
Assimilation and Dissimilation, The Classical Journal, 68, 267-271.
Schwyzer E. (1956). Griechische Grammatik. Mnchen: Beck. Shipp,
G.P., (1958). The phonology of Modern Greek. Glotta 37, 233-58.
Steriade, D., (1987). Redundant values. Chicago Linguistic Society
23, 339-62. Straka, G., (1979). Contribution la description et
l'histoire des consonnes L. In G. Straka (ed.), Les sons et les
mots, Paris: Klincksieck, 363-422. Suzuki, K., (1998). A
typological investigation of dissimilation. Unpublished PhD thesis,
University of Arizona. Available at ROA. Threatte, L., (1994). The
Grammar of Attic Inscriptions. Vol. I, Phonology. Berlin: De
Gruyter. Toufexis, (forthocoming). Creating a database for the
Grammar of Medieval Greek project. In I. Mavromatis (ed.),
Proceedings of the International Conference Neograeca Medii Aevi,
University of Ioannina, September 2005 Tsirpanlis, Z.N., (1982).
(17-19 .). , 8 (1982), 7-21. Vranousi, E.L., (1980). . . 1: ,
Athens 1980.
306