Auxiliary Verbs and the Participle in the Tsakonian Dialect: Towards

a Periphrastic Verbal System

Nikos Liosis
Institute of Modern Greek Studies, Aristotle University of Thessaloniki
1 Introduction
One of the many aspects that mark out Tsakonian from the other Modern Greek
dialects is the way it organizes the morphology of the verb. The most striking
characteristic is the periphrastic nature of the indicative system, which has been
remarked upon in all the classic works on the dialect, such as those by Pernot
(1934), Deffner (1880, 1881), (1923) and (1951, 1999), and
has also been discussed by researchers who have made particular studies of
periphrasis in Greek, such as Aerts (1965), who devotes several pages of his work
Periphrastica (84, 102-9, 125-7) to a description of the unusual way that
Tsakonian forms the present and perfect tenses.
2 Formation of the periphrases
The present and imperfect (henceforth referred to as imperfective periphrases) are
formed in both voices using the auxiliary (in the present and the imperfect
respectively) and the present participle, e.g. emi rfu (Peloponnesian Tsakonian
(Pel.T.)) / rfo ma (Propontis Tsakonian (Prop.T)) (= * ) , ema rfu
(Pel.T) / rfo ma(ni) (Prop.T) (= * ), emi rafmene (Pel.T) /
rafmne ma (Prop.T) (= * ), ema rafmene (Pel.T) / rafmne
ma(ni) (Prop.T) (= * ). The present perfect and past perfect
(henceforth referred to as perfect periphrases) are formed in the active voice using
the periphrastic auxiliary verb (in the present and the imperfect respectively)
and the verbal adjective in -, e.g. emi xu raft (Pel.T) / xo ma raft
(Prop.T) (= *[ ] ), ema xu raft (Pel.T) / xo ma(ni) raft (=
*[ ] ) and in the passive voice with the monolectic auxiliary
verb (in the present and the imperfect respectively) and either the verbal
adjective (in Pel.T and Prop.T) or the passive perfect participle (in Prop.T only),
emi raft (Pel.T) / raft ma (Prop.T) / ramne ma (Prop.T) (= * /
), ema raft (Pel.T) / raft ma(ni) (Prop.T) / ramne ma(ni) (Prop.T)
(= * / ).
Leaving to one side the similarly periphrastic future tense, the modal origin of
which marks it out from the other tenses as regards both its usage and its
morphosyntactic formation (for the future periphrasis in Tsakonian, see
2010), the only instance in Pel.T where we find synchronically monolectic
indicative types is the aorist paradigm, e.g. erva, ervere, erve etc. I, you,
he/she/it wrote etc. But in the now extinct Tsakonian dialect of the Propontis, just

469

as frequently as the monolectic aorist types we find periphrastic structures with

the same function (henceforth referred to as perfective periphrasis), formed using
the present (or imperfect) of and a type of participle which is generally
considered either to be derived from an older active perfect (Deffner 1880) or
thematic aorist, or else to imitate the ending of the present participle of oxytone
verbs (Pernot 1934:240-1), e.g. aapik ma (= * ) I loved. The
perfective periphrasis is unknown in todays Tsakonian dialects of the
Peloponnese ( 2007:783-6); similar fossilized structures have, however,
been sporadically recorded by 19th and early 20th century researchers, e.g. emi
orak (= * ) I saw, emi apostak (= * ) I opened
my legs.
3 The participles
Each of these periphrases is historically linked to the development and
functionality of the participle with which it is formed. A historical examination of
these participles may therefore lead us to useful conclusions, or at least to useful
questions concerning the history of periphrasis in this dialect.
In the early mediaeval period, the only participles that can generally be
considered to have remained a living part of the Greek language are those which
belonged to the regular paradigms of the old second and first declensions, i.e. the
mediopassive present participles in -, -o, - and the perfect
participles in -, -, - ( 2006:435-7). In time, however, the
former died out, together with the equivalent active types, while the latter were
preserved either as adjectival determiners or as predicates in stative periphrases of
the perfect together with the verb , e.g. am written. It is
easy to see that Tsakonian had already diverged from the situation described here.
The biggest difference is that the active participles and the passive present
participle not only did not disappear, but have preserved many of their older
functions and even developed new ones. At the same time, rather amusingly, the
verbal adjective has taken the place of the passive perfect participle, the only one
to have been preserved in SMG and in most of the other Modern Greek dialects.
3.1 The present participles
Let us take a closer look at the present participle. In the other Greek varieties, the
only relic of the old active participles of the third declension with their
complicated declensional paradigm is the indeclinable participle in -(),
which has been preserved only as an adverbial determiner of the subject, e.g.
he/she left crying. Conversely, in Tsakonian the participle is
genuine: In the active voice there are 4 morphologically distinct types, which
make distinctions of gender and number, e.g. rf-u /-o (masc., sing.), rf-a
(fem., sing.), rf-unda /-(o)nda (neut., sing. & plur.), rf-unde /-(o)nde (masc. &
fem., plur.). The passive present participle in -mene / -mne retains the same
degree of finiteness: here too there are 4 different types, e.g. raf-mene / -mne

470

(masc. & neut., sing.), raf-umna / -mna (fem., sing.), rafumni / -mni (masc.
& fem., plur.), raf-mena / -mna (neut., plur.).
The absence of case distinctions and the convergence of masculine and
feminine in the plural of both voices and masculine and neuter in the passive
singular should not be considered symptomatic of a process of grammaticalization
analogous with that which ultimately produced the SMG type in -, since
these exact characteristics are common to all determiners in this dialect, compare,
for example, the following adjectives and pronouns, e.g. kuvni black (masc. &
fem., plur.), at big (masc. & neut., sing.), ndei they (masc. & fem.), pe
who/which (masc. & neut., sing.). On the other hand, the neuter ending -unda / ()nda (in place of the expected *-u < -), which is historically derived from
either the masculine accusative or the neuter plural, shows that the changes which
ultimately produced the Modern Greek indeclinable type had begun in Tsakonian
too, but stopped at a very early stage, presumably supported by the other
participles, which had not been affected by these changes.
The present participle preserves to this day a high degree of functionality. A
characteristic example is the predicative use of adjectival participles which have
been replaced in Modern Greek by complement clauses. These participles agree
with the subject or the object of verbs which belong to one of the following
semantic fields:
a) Verbs of feeling, perception, knowledge (agreement with the object), e.g. li
eci ora san traj, li san nu xunde she saw some running like billy-goats,
others like donkeys (Pel. T.; 1986, 3:414), orka to spre s are ni
rxno I saw that he was scattering his seed sparingly (Prop. T; 1986,
1:125), s eka jonde I heard them laughing (Pel.T. 1999:141),
erka akainmene tas ton krpo I found him wallowing in dung (Pel.T.
1999:141) etc.
b) Verbs of starting, finishing, tiredness, satiation, harm; in this category are
included a large number of verbs with many synonymous meanings. It seems that
in Prop.T. this category is absent. Agreement is exclusively with the (logical)
subject, e.g. ga randu I started (= took) running ( 1986, 1:131),
apastse tsan ra a he didnt finish talking all that time ( 1986,
1:99), apombtse a rsa mi rota ti javti my tongue got tired of asking
passers-by ( 1870:48), ekfte paa na namu ri she rushed to come
and see us ( 1986, 2:85), ekserokrni i pe mi xa tshapite my
feet were chapped with running barefoot (, 2:341-2), alioe ksikzu
he went cross-eyed from looking (, 1986, 1:51) etc.
Some of these participles, more often than in SMG, are nominalized, e.g. to
aposurmene (= * ) the dregs (, 1986, 1:117), fus
(= * ) the devil ( 1986, 1:325), exmene (= * )
the rich man ( 1986, 1:307), xu (= * ) the owner, cf. the
utterance: a sembrpsi me ton xu to zovi he will go into partnership with the
owner of the pair (of oxen) ( 1986, 1:312).

471

These participles are also used very frequently to express adverbial relations
which in SMG are realized either with the indeclinable type in - or with
adverbial clauses:
a) manner, e.g. ekne agoru he came panting ( 1986, 1:16),
b) purpose, e.g. i ci im barnde nde ta cin the wolves come to eat the
goats ( 2007:536),
c) time (simultaneous actions), e.g. paru ts o mbap na kpsi kaia ja ta prata
ori tam brovta para randa while grandfather was coming to cut
branches for the sheep, he sees the ewe come running ( 2007:536),
d) opposition, e.g. embkate em lunde you did it against our will
( 1951:138).
As we can also see from this last utterance, in many cases these adverbial
participles do not agree with the subject of the verb, but may be realized as a freestanding, absolute part of the sentence (nominative absolute), e.g. ekne tn to
ci, o e mu he came in the heat of the day, the sun trembling (
1986, 3:102), ra nde eknai tse ta kambza while we were eating, the
children came too ( 1999:141).
3.2 The active past participle
The second active participle in Tsakonian appears somewhat more problematic.
From the available data about Pel.T. ( 1846, 1870,
1920:58-9, Pernot 1934:241, 1951:118) I gathered 14 instances
overall: rav (erva I wrote), orak (orka I saw), (e)kan (ekna I came),
zak (ezka I went), (e)mbik (embka I did), vai (evaa I yelled), ak
(eka I burned), ark (erka I beat), lev (elva I peeled), pek (epka I
said), erak (eerka I got old), ak (eka I heard), ferk (frka I
brought), apostak (apostka I opened my legs).
With regard to morphology, the same applies as in the case of the participles
discussed above: here too are preserved 4 different suffixes of gender and
number, e.g. rav- /- (masc., sing.) ~ rav-a /- (fem., sing.) ~ rav-nda /(ta) (neut., sing. & plur.) ~ rav-nde /-te (masc. & fem., plur.). The stem of
these particles must be considered to derive historically from a perfect form.
Support for this view can be easily found in the case of those participles
containing the element -k-, e.g. ork-a < , and also for those with stems in
a voiced fricative, which (to appear) has shown to be derived from the
stems of attested Doric perfect forms, e.g. (e)rv-a < Dor. - (= Att.
-), tar I agitated < *tar-a < Dor. - (= Att. -). It is
therefore logical to assume that the endings of these participles are also derived
from the old endings -, -, - of the active perfect. The objection raised by
Pernot (240), who was unaware of the Asia Minor subdialect, that the feminine
ending, according to the phonetic laws of the dialect, should have had the form za, was rightly rejected by Aerts (1965:84, footnote 1), since it could have arisen
from the (early medieval) ending -. A similar conversion is shown by the

472

feminine ending - in Prop.T. Another important point is that these participles,

like the equivalent perfect indicative, have undergone syncretism with the aorist
and were interpreted as perfective. The same process took place in all forms of
Greek: The far-reaching syncretism of aorist and perfect forms also includes the
participles (Aerts 1965:90). The alternation between aorists in - and - was
therefore a universal characteristic of mediaeval Greek, referred to in all the
bibliography on the subject (e.g. 2006, 443); in Tsakonian, however, for
the perfective stem of the indicative, the prevalence of the element -k- (or, better,
the absence of the element -) is absolute. Within this framework, the only
example that appears to continue a thematic aorist, the type (e)kan- /- (Aor.
ekn-a), does not constitute a real exception. Even if we accept the fact that some
role was played by the thematic aorist participle endings -, -, -, the most
important point is that what ultimately developed was an active perfective
participle opposed to the imperfective participle of the same voice.
It is, however, worth noting that in the subdialect where it was best preserved,
that of the Propontis, I have been unable to find a single example of its use
outside of perfective periphrases. In Pel.T too, the only two examples where it is
used adverbially come from (1920:59): kan tse orak o adra aortse tan
mbele having come and seen it, Andreas bought the vineyard and ravnda tu
kambzune, apolkame ton karp when the children wrote, we sent the wheat. In
both cases, the participles describe a previous action. In the second example, not
only is the participle freely, but its subject appears in the genitive (a genitive
absolute), a fact that casts doubt on the authenticity of the examples recorded by
the Tsakonian man of letters.
4 Interpretation
The necessary condition which allowed the formation and preservation of
periphrases in Tsakonian was the fact that the participles had remained very much
a living part of the morphology of the verb. I hope that the discussion presented
above has made this clear. Aerts (1965:98) observes this in reverse with regard to
the majority of Greek dialects It goes without saying that where the participle
has disappeared, the periphrases have also been dropped. In other words, in
contrast to other varieties of Greek, Tsakonian has diachronically had recourse to
a cross-linguistically useful and convenient tool for the construction of
periphrases. Furthermore, the use of participles added another semantic
characteristic to the verbal stem, that of gender, and it seems likely that the need
to achieve gender distinction was a factor in the preservation and generalization of
periphrases in the Tsakonian verbal paradigm, although it is admittedly difficult to
determine which came first. The opposite development in Modern Greek, i.e. that
pressure in favour of a lack of gender marking on the verb could have played
some role in the replacement of perfect periphrases with participles by those using
the infinitive, has not been taken into account in the bibliography.
The fact that something is possible in a language does not, of course, mean that
it is inevitable. Garcias (1987, in Bybee, Perkins and Pagliuca 1994:133)

473

statement that a periphrasis is born when there is a need to express a specialized

or new meaning that is not expressed (or not overtly expressed) by the monolectic
type, is of course correct; it is, however, too general to explain why Tsakonian is
the only Greek dialect to have developed an imperfective suppletive periphrasis
(for the term suppletive periphrasis see Aerts 1965:3), i.e. why it has replaced the
monolectic present and imperfect types with a periphrasis which has taken on the
full range of meanings of the tenses in question. This periphrasis does not
function simply as an alternative possibility, as is the case, for example, with the
pair - in Ancient Greek, nor does it have an
exclusively progressive meaning, as in the case of the English I am writing (in
contrast to I write), but constitutes a fully functional imperfective stem, and
consequentially may admit equally of habitual or continuous and progressive or
non-progressive readings.
This fact also provides the most important argument in favour of the
hypothesis that the periphrastic present and imperfect constitutes an innovation of
the Tsakonian dialect, rather than continuing a tendency inherited from Ancient
Greek. According to Aerts (1965:17), the present periphrasis in Ancient Greek is
predominantly situation-fixing, situation-describing and intransitive, while later
a new form of application appears, namely that which we shall call the
progressive periphrasis, after Bjrck. This progressive periphrasis is confined to
religious texts, and therefore cannot have been the model for the construction of
the periphrasis in Tsakonian, the speakers of which did not convert to Christianity
until a much later date (108-9). The author also rejects the hypothesis that we are
dealing with a Laconism (106): A study of the inscriptions of Laconia [...] has
shown that the few periphrases that occur are all of the type. The
Tsakonian periphrasis is, therefore, almost certainly not a Laconism.
Aerts (109) ultimately concludes that this is an independent development in
Tsakonian which was influenced by the perfective periphrases of both voices and
arose as a mechanism for differentiation between the indicative and subjunctive
moods, the latter of which, in the active voice, coincides exactly with a
reconstructed monolectic indicative and is combined with the subjunctive marker
, e.g. eni rfu ~ na rfi (= * ~ ). The view that the final
push for the development of the new structure was given by the tendency of this
dialect to make distinctions of mood, more specifically the basic distinction
between indicative and non-indicative (subjunctive), with the greatest possible
morphological facility, is discussed, not without reservations, by (to
appear) too (see below). also considers the fact that the old monolectic
types of the present indicative are not lost, but instead re-used in the subjunctive,
to be a characteristic example of the change of a verbal system through role
redistribution among the existing types.
I agree that the perfect periphrases lent support to the formation of the present
and imperfect periphrases, although the fact that the verbal adjective is used
instead of the participle in the passive voice weakens this argument to some
degree. I do however believe that the need which caused the use of periphrasis to

474

become general in these two tenses was the tendency for overt morphological
distinction of aspect (imperfective vs. perfective), and not modality (indicative vs.
subjunctive). Because if we accept Aertss view, it is difficult to explain why in
the aorist the indicative has remained monolectic; also the presence of the marker
would appear to be redundant. In addition, in the imperfect, where the
distinction between indicative and subjunctive is meaningless, there should be no
need for replacement of the monolectic type. (to appear) points out that
the opposition between aorist (perfective) and imperfect (imperfective) was better
expressed through the replacement of the monolectic imperfect by a periphrastic
one. And only then this construction was analogically spread to the present tense.
That this is the right order is obvious in other dialects where the need to underline
the imperfectivity of the imperfect lead to its augmentation with frequentative
suffixes, as in I cry ~ (Pharasa, Asia Minor; Dawkins 1916: 180),
in I love ~ (Paxoi, Ionian Islands; & (to appear),
in drink ~ -- (Silli, Asia Minor; 1968:81) etc.
This general explanation could be satisfactory for our case, because it connects
tendencies observed in many Modern Greek dialects. However, I think that in
Tsakonian in particular there are additional phonetic and morphological reasons to
believe that the present and imperfect periphrases took shape under pressure for
clear expression of the bipolar opposition between perfective and imperfective: as
well as the undoubtedly perfective indicative stems in -k-, the dialect also makes
use of the perfective element -s- (from the old sigmatic aorists), exclusively for
the formation of the perfective subjunctive, e.g. ork-a I saw ~ *na (o)rs-u,
ek-a I slaughtered ~ na *s-u, which according to the phonetic laws of the
dialect was subsequently lost between vowels (> na ru, na u), i.e. in oxytone
verbs and those where the stem ends in a vowel. However, this development
meant that the stem of the perfective subjunctive now coincided with that of the
monolectic present indicative. And how better to solve this problem than to
replace the stem of the latter with a periphrasis that uses the present participle and
is therefore characteristically marked as imperfective? Schematically (the
periphrases are in brackets):
mood

indicative subjunctive
indicative subjunctive
[eni u] (na) -n-u
imperfective *-u
(na) -u
perfective
e-k-a
(na) *s-u > (na) -u
e-k-a
(na) -u
aspect

Furthermore, the coincidence of the perfective and imperfective subjunctive is

remedied by the addition of the frequentative affix -n-. This means that we have 4
clearly distinct stems in a symmetrical system: of the old perfect (ik-), of the old
aorist (i-), an extended n-stem (in-) and a periphrastic one (Aux. + i-).
However, if the distinction between perfective and imperfective aspect is
expressed in this dialect by the distinction between monolectic types and
periphrases, should not the imperfective subjunctive be periphrastic too? The

475

answer may be found in the dialect of the Propontis. Examples such as those
given below are abundant in the available material: m sa ni kakomelet to kav
dont bring bad luck on the child ( 1986, 2:12), m sa fozmna man
dont be afraid, mother ( 1986, 2:221), m sa lo pi klvo ma dont
say that I steal ( 1986, 2:79), m sa sinorizmne dont try to get even
( 1986, 3:206) etc.
It is clear that we are dealing here with periphrastic imperfective subjunctives,
which appear, however, only in a negative context, that is together with the
negative particle of the subjunctive, and in the second person. This is not by
chance; according to Veloudis (1987), second-person negative structures are, as a
general rule in Greek, the only ones that can appear without the marker . The
only instance of an affirmative periphrasis (where, obviously, is present) is the
utterance na sa x(o) tan galosna t may you have your health ( 1986,
2:26) which does not, however, constitute a genuine exception because the verb
lacks the distinction between perfective and imperfective stems. Thus for the
2nd person the equivalent schema in Prop.T is:
mood
aspect

indicative subjunctive

imperfective [rfo sa]

perfective
rve

(mi) [sa rfo]

(na) rps

I believe that the morphosyntactic regularity of this phenomenon allows us to

interpret it as systematic, rather than as an example of language attrition (for
examples of the latter see 2007:686-839). In fact, Givon (1979) numbers
negatives among the most conservative contexts, where old structures are most
likely to be preserved for the longest time. Another relevant archaic aspect of
Prop.T is that it does not make use of the extension -n- in the formation of the
imperfective subjunctive stem: in other words this subdialect is deprived of the
other means to reinforce imperfectivity. Is it perhaps possible that the presence of
the subjunctive marker led to the final correction of this mood, contributing to
the prevalence of the monolectic types?
If the above description is correct, it is another indication of the increased
importance of aspect in the Tsakonian verbal system: the periphrases fill the gap
that opened up in place of the imperfective stem, restoring, in Haspelmaths
(2000:656) terminology, the paradigm symmetry of the verb; given that the
imperfective and perfective stems of verbs whose stems end in a consonant never
coincided (e.g. raf- / raps-) obviously the periphrases of these verbs were
formed analogically with those of oxytone verbs and those whose stem ends in a
vowel we must hypothesize that the criterion operating here is likely to be that
of inflectional generality (for the term see Haspelmath 2000:656), i.e. the need
to generalize in the latter two classes of verbs too a morphologically transparent,
declinable type for the imperfective stem.

476

Another possible indication that the imperfective was identified with

periphrastic forms and the perfective with bound expressions is the presence in
Prop.T of periphrastic imperfective imperatives, e.g. tshaupta sa take your shoes
off (imp.) ( 1986, 3:260-1; cf. the monolectic perfective imperative
tshauptos). In Pel.T the distinction was once again realized by the addition of -n-,
e.g. rine see! (imp.)~ ra (perf.).
5 The perfect periphrases
Moser (2008) refers to a similar reorganization of the Greek verbal system based
on the bipolar opposition perfective vs. imperfective, which resulted in the
absence of the perfect tenses from the system for several centuries. When the need
arose again, the only available forms were the periphrastic ones. We may suppose
that in Tsakonian the gap in question was originally filled by the periphrases with
the active perfect participle. Soon, however, these periphrases converged, in terms
of both morphology and function, with the aorist, as we can see from examples
such as the following (Prop.T), where any (past) perfect reading of the periphrases
are not accepted:
a) utterances with temporal determiners, e.g. mema arev tar o jatr At once
(suddenly) the doctor became angry ( 1986, 1:21),
b) narratives where the periphrases are linked by parataxis with monolectic aorists
having a synonymous meaning, e.g. pondie, kriak tse pentse he caught cold,
he fell ill and died ( 1986, 2:86),
c) instances of alternation between direct and reported speech within the
framework of a narrative, e.g. po tinen itom n nda, pek ta o pap theres
no escaping from him, said the priest ( 1986, 1:236),
d) temporal modifiers, e.g. ast tar etne, ma t akuk ttoa oj he went out of
his mind when he heard such things ( 1986, 3:133).
It was therefore necessary to resort to another periphrasis, which was
ultimately constructed using the old verbal adjectives in -, e.g. xo ma trijit t
amb I have harvested the vineyard (Prop.T.; 1986, 3:253), e xu
eit tan eja I have tied up the goat (Pel.T; Pernot 1934:225) etc. To the
question of why, with the partial exception of Prop.T, the dialect did not make use
of the passive participle in -, a possible answer is that this would have
created confusion with paroxytone types of the present participle, e.g. rafu-mna
being written (fem., sing.), rafu-mni (masc./fem. plur.).
6 Conclusions
In previous papers I have spoken about the Tsakonian future and counterfactual
periphrases ( 2010, Liosis 2010) which are formed using the auxiliary
. If we also consider the generalized use of this auxiliary, together with ,
in the formation of imperfective, perfective and perfect periphrases in this dialect,
we can comprehend the central role it played in the system as this was
reorganized in order to achieve new distinctions of tense, aspect and modality, or
to include new semantic characteristics such as gender. The emergence and

477

prevalence of periphrases in the dialect must therefore be seen from this

perspective. Dahl (1985), Bybee & Dahl (1989:56), Bybee & Perkins & W.
Pagliuca (1994: 104-124) say that, cross-linguistically, periphrases are usually
associated with perfect and progressive, while past, perfective and imperfective
are usually expressed monolectically. Thus the past, perfective and imperfective
periphrases in Tsakonian show that it is not a usual language, and this fact is
likely to have consequences for the theory of grammaticalization itself.

References
Aerts, J. 1965. Periphrastica. Amsterdam: Hakkert.
, . 1987. * , , (*) ,... / ,
, , ... : . Studies in
Greek Linguistics 8. Thessaloniki. 293-309.
Bybbe, J. and . Dahl. 1989. The creation of tense and aspect systems in the
languages of the world. Studies in Language 13. 51-103.
Bybee, J., R. Perkins & W. Pagliuca. 1994. The evolution of grammar: tense,
aspect and modality in the languages of the world. Chicago: U.C.P.
Dahl, . 1985. Tense and aspect systems. Oxford: Blackwell.
Dawkins, R. 1916. Modern Greek in Asia Minor. A study of dialect of
Silly, Cappadocia and Pharasa. Cambridge: Cambridge University Press.
Deffner, M. 1880. Das zakonische Verbum und seine Formen. Archiv fr mittelund neugriechische Philologie 1. 77-87, 167-185.
Deffner, M. 1881. Zakonische Grammatik. Berlin: Weidman.
, . 1923. . : .
Givon, T. 1979. On understanding grammar. New York: Academic Press.
Haspelmath, M. 2000. Perifrasis. In Morphologie: Ein internationales Handbuch
zur Flexion und Wortbildung, ed. by G. Booij, V. Lehmann & J. Mugdan, 655664. Berlin: DeGruyter.
, . and . to appear in 2011. . In
, ed. by . . :
.
, . 1951. . .
. . . :
Institut Franais dAthnes.
, . 1968. . :
.
, . 1986. . vol. 1-3. :
.
, . 1999. (
). 15.
, . 1920. . .

478

, . 2007. . PhD.
Aristotle University of Thessaloniki.
, . 2010.
: . Studies in Greek
Linguistics 8. Thessaloniki. 372-384.
Liosis, N. 2010. Counterfactuals in Tsakonian: a contribution to the history of
and . In On-line Proc. 4th International Conference of Modern
Greek Dialects and Linguistic Theory. Chios. 89-102.
Moser, A. 2008. The changing relationship of tense and aspect in the history of
Greek. Language Typology and Universals 61.5 -18.
, . 1846. () . .
, . 1870. . .
Pernot, U. 1934. Introduction l'tude du dialecte Tsakonien. Paris: Les Belles
Lettres.
, . to appear in 2011. . In
, ed. by . . :
.
, . 2006. . .
Introduction-translation in Greek by . and . . .
.

479