Palmer L.R. The Greek Language
Подождите немного. Документ загружается.
< previous page page_49 next page >
Page 49
dative, on the evidence of the Arcadian duals like , may well have been *dwoun, and possibly -phi was
added to form the instrumental or locative;1 alternatively the stem form may have been *dwou-. For 'three' there is
only the dative tirisi = trisi. A compositional tri- occurs in tirijowe = triowes 'with three handles' and tiripo =
tripos 'tripod cauldron' There are also compositional forms for 'four' (qetoropopi = qwetro-popphi * instr. plur.
'quadrupeds', qetorowe = qwetrowes* 'with four handles'), 'six' (wepeza = wespegja <
) and 'nine'
(enewopeza = en(n)ewopegja 'with nine feet').
The Verb
The augment is absent with the one exception apedoke, as against apudoke, doke = doke, dekasato = deksato,
epidato = epidasto, qirijato = qwriato*, teke = theke, wide = wide, zeto = skheto or gento (?). Reduplication
shows no peculiarities in the perfect: dedemeno = dedemeno, dedomena = dedomena, epidedato = epidedastoi,
eeto = (h)ehentoi 'have been sent'. Examples in the present are didosi = didonsi, didoto = didontoi; the personal
name pepi-temenojo = pepithmenojo may be either aorist or perfect middle participle of
.
The third person singular middle endings, singular and plural, of the present and perfect are written -to, and this is
interpreted as -(n)toi, which recurs in Arcado-Cypriot (see pp. 298f.): euketo = eukhetoi, (o)uruto = wruntoi,
epidedato = epidedastoi, qeqinoto = gwegwinotoi* 'it is carved (with patterns)', -didoto = dido(n)toi 'they are
given'.
The thematic verbs in the active appear to have the normal endings: eke = ekhei, ekosi = ekhonsi, operosi =
ophelonsi/ophel-lonsi. The athematic verbs show the usual Ablaut alternation: kitijesi = ktiensi (< *ktei-/kti-), later
. There is a verb ije- = hie- (*ise-/ ) 'sacrifice, etc.', homophonous with hie-'send' ( < *jije-); known forms
are (jo)ijesi = hiensi, ijeto (which may be aorist middle) and apieke = aph-ieke 'he made offering' (later
).
The verb tereja third person singular (with infinitive terejae, see below) is an athematic vowel stem like Hom.
.
Future formations are dose = dosei, (jo)dososi = dosonsi, ze-
1Cf. hikwoiin-phi above.
< previous page page_49 next page >
< previous page page_50 next page >
Page 50
someno = gjessomenoi 'for him to boil'. Of great interest is ewepeseomena = ewepsesomena * 'to be woven', a
future with the expected e-grade < *H1webh-, a verb replaced later by the denominative
( < A
parallel formation with suffix -se is asesonsi = assesonsi 'they are to fatten up' < *H2es-'sate with liquid food'.
Another type of future (p. 311) is demeote = demehontes 'going to build'.
Active aorist formations are doke = doke, teke = theke, apieke (see above), (jo)oporo = ophlon 'they owed' or 'I
owed'. Strangely, there are no certain sigmatic forms1 (but see below on erase). Sigmatic Middle forms are dasato
= dassato, and dekasato = deksato; other types are qirijato = qwriato* 'bought', and possibly zeto = gento or
skhento 'get'.
The following finite forms of the perfect can be quoted: epidedato = epidedastoi, qeqinoto = gwegwinotoi*
(denominative verb from *gwinos* 'a graving tool', see above), eeto = (h)ehentoi 'have been sent'.
The thematic infinitive shows the uncontracted form -ee = -ehen: ekee = ekhehen, anakee = anagehen, wozee -
worgjehen, eree = erehen 'to row' (primary verb < *H1erH1-, later replaced by the denominative
). The
form terejae, alternating with the indicative tereja (see above), is an athematic vowel stem; a similar type of
infinitive occurs in Lesbian
, etc. (see below).
Participial formations are frequent. Present' active ekote = ekhontes, ijote = iontes, operota = ophelonta/ophellonta,
operosa = ophelonsa/ophellonsa, wozo/wozote = worgjon/worgjontes; middle reqomeno = leiqwomenoi*,
toroqejomeno = stroqwheiomenos*, kitimena = ktimena, oromeno = horomenos 'watching', wozomeno =
worgjomeno 'under construction'. Future: zesomeno = gjessomenoi, ewepeseso-mena, demeote (see above). Aorist:
active akera2te = agerjantes;2 middle in numerous personal names kurumeno = Klumenos, pirameno =
Philamenos, etc. Perfect: in the active the sigmatic suffix *-wos* appears throughout and there is no trace of -wot-:
araruwoa = ararwo(h)a, araruja = araruia 'joined', tetukowowoa2 = tetukhwoha 'finished', keketuwoe =
kekethwohes (? <
1 Perhaps the aorist interpretation of ereuterose = eleutherose, 'he has excused' is to be preferred to the fut.
eleutherosei.
2 The spelling with ra2 suggests that -rs in ager-sa- had developed to a palatalized -rr*.
< previous page page_50 next page >
< previous page page_51 next page >
Page 51
'help'). The personal name widowoijo = Widwohios is derived from widwos *, the masculine perfect participle of
weid-/wid-'know', with zero grade of the root. The form dedikuja = dedi-kuiai 'having finished their
apprenticeship' (the etymological connections of the root dik- are obscure; an error for dedi<ka>-kuja is
suspected). Middle forms are frequent: dedemeno = alede-meno 'bound', kekaumeno = kekaumenos 'burnt',
[pe]pu2temeno = pephut(e)meno 'planted', araromotemena = ararmot(e)mena 'fitted out' (others interpret
ararmotmena), ajameno = aiaimeno 'inlaid' (?) (there is an agent noun a3tere = aiteres). Augment for re-
duplication appears in erapemena = erraphmena 'sewn'.
The following verbal adjectives in -to- occur: akitito = aktitos, aneta = an(h)eta 'remitted', kakodeta = khalkodeta
'bound (riveted) with bronze'. Surprising is what appears to be a gerundive form qeteo/qetejo/qetea2/qetea
contrasting with ono and so presumably a transaction term. Cypriot evidence (tesata =
) favours Ventris's
first interpretation1 gwheste(ja)* (< *gwhedh-*, 'miss', 'long for', see p. 235) rather than qweite(j)a* 'to be paid',
see p. 314 for the gerundive formation).
The following forms of the verb 'to be' are found: eesi = ehensi 'they are', eo/eote = ehon, ehontes ptcp., and
perhaps esoto = essontoi 'they will be' and -ape = apes 'he was absent'. The alleged eni = enesti is a proper name.
Possible modal forms are obscured by the ambiguous spelling rules. If ekeqe is a prospective form with the particle
-qe distinct from -qe 'and' (which makes no sense in the context, see below), then eke may stand for subjunctive
ekhe. The same may be true of toe = thoe 'he is to pay'. Another possible example is oqe erase, the final item in a
list of 'missing' herdsmen and animals: this is open to interpretation as hoqwe(n)* elase (the man and his animals
will be absent) 'when he drives'.2 For the further connections of the particle qe distinct from -qwe* 'and', see pp. 67
f. Of relevance to the phonological history is the fact that a combination with the negative ouqe also occurs, which
cannot be translated 'and not', 'nor'. Thus ouqe akerese is also likely to be a
1 This is also favoured by contextual analysis, yielding a meaning 'not issued, to be demanded'.
2 But hoqwe* (cf. ) is equally possible: this would be an instrumental form equatable with Thessalian
(and possibly also with Cypriot ) and hence of interest for the affinities of Linear B Mycenaean.
< previous page page_51 next page >
< previous page page_52 next page >
Page 52
modal form 'he will not...'. Note that in this combination the rule relating to the development of the labiovelar has
not applied (see above on qoukoro, suqota).
Important for the assessment of dialect relations is the construction of the preposition paro. Expressions like eke
onato paro...pomene are most easily translated 'he has a lease (?) from the shepherd' (pomenei, dative), but a sense
'chez the shepherd' has been argued. Decisive is PY Un 138 qetea2 paro dunijo if the gerundive is identified as
gwhesteha * 'to be demanded (from D.)'
Miscellaneous
Patronymic adjectives are used instead of nouns in the genitive' etewokereweijo = Etewoklewehios 'Son of
Eteocles'. The material adjectives alternate between -ejo/a- and -ijo/a-: raeja = laheia 'stone', weweea =
werwe(h)e(h)a 'woollen', kakejapi/kakijo = khalkeiaphi/khalkio 'of bronze', etc. Adjectives in -went- are frequent,
and the feminine has the analogical full grade -wessa (for the expected
> *-wassa). The suffix is added
directly to the stem-consonant: pedewesa = pedwessa (with zero grade ped-of pod-), toqidewesa = torqwidwessa*
'with spirals' (?), odatuweta = odatwenta 'having teeth' (of wheels) ( < *odnt-went*, with zero grade of root), etc.
A strange form is potinijawe(i)jo 'belonging to Potnia', perhaps with a suffix -weio- parallel with the -
meio-attested in
.
Prepositions (verbal prefixes) worthy of comment are ano (only in compounds like anoqota, anomede), which
possibly owes the final vowel to its opposite ; final o occurs also in paro (elsewhere para); apu (apu
kekaumeno, apudosi, etc.); peda (peda watu = peda wastu); opi 'on', 'over', 'in the charge of', 'at the head of' (also
in compounds like opoqo = opoqwois* 'with blinkers', opidarnijo 'man in charge of, at the head of, the damos',
opia2ra = opihala; cf. later
). The word posi may stand either for posi or porsi (see below); it is used
adverbially as in ouqe anija posi 'nor are the reins attached'.
A surprising omission from the Linear B texts is any form corresponding to
/ 'and', though enclitic -qe is
common. The negation ou is always written as a proclitic.
< previous page page_52 next page >
< previous page page_53 next page >
Page 53
4. Dialect Variation and Chronology
The Mycenaean language, attested in so many widely separated places, is of a striking uniformity. The only
differentiations of any note are 1, the occasional dative in -i instead of-ei and 2, the occasional representation of *-
m * as -a- instead of -o- even after a labial consonant. These aberrations from 'normal Mycenaean' are confined to
a few scribes and hardly deserve the positing of a 'special Mycenaean' dialect. Such uniformity suggests that
'Linear B' is a 'chancellery language', though it was doubtless originally based on a living dialect, the affinities of
which will be discussed below. Such a fossilized language will surely have been more archaic than the spoken
language of the scribes, and occasionally this will have intruded into their written work. At all events the dative in
-i had superseded -ei by the time of 'alphabetic' Greek except for survivals in proper names.
The uniformity of language (and administrative practices) would be all the more impressive if it was extended not
merely in space but in time, for it was long believed that the Knossos tablets (c. 1400 B.C.) were some 200 years
earlier than those of the Mainland (c. 1200 B.C.). The evidence for the centuries-long unity and tenacity of such an
administrative language deserves the attention of philologists. However, in 1958 C. W. Blegen, the excavator of
Pylos, questioned the chronology of the Knossos tablets and suggested 1, that the Throne Room at Knossos was
inserted in the 'Minoan' Palace in the LM III B period and 2, that the Linear B tablets belonged to the end of that
period. (Here it is relevant to recall that the tablets were sun-dried and would soon have disintegrated if they had
not been accidentally baked in the fire that accompanied the final destruction.) A point that had escaped notice
gave strong support to the thesis: Evans and his later assistant J. D. S. Pendlebury both stated that the East Wing
(the Domestic Quarter) had been cleared of debris after the destruction of c. 1400 B.C. and reoccupied in the
subsequent LM III B period. Evidently the deposit found by the excavators must in that case have belonged to this
'Reoccupation'; it contained large numbers of Linear B tablets, giving cross-references to deposits in other parts of
the palace. This 'unity of the archives' was later reinforced by a study of the scribal 'hands'.
< previous page page_53 next page >
< previous page page_54 next page >
Page 54
Blegen's thesis focused attention 1, on Evans's evidence for dating the construction of the Throne Room and 2, on
the evidence submitted bearing on the chronology of the tablets. It is now established 1, that the sub-floor sherd
material referred to, but not published, by Evans was of LM III A 2/III B type, which supports the LM III B
construction date, and 2, that the elaborate stratigraphy presented by Evans as 'decisive' for the LM II date of the
tablets was wholly invented by him to support his 'Reoccupation' theory. The tablets and pots assigned by him to
the room in question and sited on different floor levels had been found in different widely-separated parts of the
palace, while the three floor levels had also not been observed in the said room. Evans, however, stated that tablets
found in the Little Palace belonged to the later Reoccupation Period. This has been confirmed by recent
excavations in the adjacent area which brought to light an inscribed stirrup jar of LM III B date, a fact which must
be interpreted in the light of the view shared by virtually all Linear B scholars that use of the script implies a
Palace administration. Furthermore, excavations at Khania have also produced inscribed jars together with
fragments of Linear B tablets, and it has been shown that Khania was the source of some of the inscribed jars
found at Thebes. In general, it is now firmly established that Crete, so far from lapsing into isolation and poverty
after the alleged destruction of Knossos c. 1400 B.C., remained literate and an exporter to the Mainland palaces
until shortly before their destruction c. 1200 B.C.
Philologists, therefore, have no solid archaeological grounds for believing that a fossilized administrative language
was in use in the Mycenaean world for something like two centuries. The new results will also have their impact on
ideas about the genesis of the Linear B script in Crete as early as the fifteenth century B.C. As mentioned above, it
was evolved in all probability from the 'Aegean' parent script independently of Linear A. The latter is
overwhelmingly a Cretan script, whereas Linear B is widespread in the Mycenaean Greek mainland in the LM III
period. Since it is attested also in Crete only during the same period,1 the simplest hypothesis is that Linear B
1 Not one single piece can be securely assigned to any phase of the Palace of Knossos earlier than the 'Last
Palace', which recent archaeological work has shown to be of thirteenth-century construction and
destruction.
< previous page page_54 next page >
< previous page page_55 next page >
Page 55
evolved on the Greek mainland independently of Linear A and was introduced into Crete when the Mycenaean
Greeks took possession of the island at some time after the beginning of the LM III B period. Considerable
numbers of sherds of this ceramic phase, many showing signs of severe burning, were found in the sub-floor
deposits of the Last Palace.
Another historical fact of importance to the philologist will be more conveniently discussed in the next chapter, but
it deserves mention in the present connection. In post-Mycenaean times Crete was occupied by Dorian speakers,
but the presence of Mycenaean elements in their dialect suggests that there was no gap between the Mycenaean and
Dorian occupation. The evidence for continuity is further supported by tribal names. Peculiar to Dorian states is the
organization into three tribes, Hylleis, Dymanes and Pamphyloi. These appear also in Crete, but there are in
addition names attributed by historians to the pre-Dorian settlers. Units with such names are believed to have
comprised elements of the Mycenaean population who were subjugated by the Dorians. One of these is Aithaleis; it
is interesting that a locality in the Pylos Linear B tablets has the name a3tareusi = Aithaleusi, dative plural of
Aithalewes, which must have been a tribal name meaning 'people of Aithalos'. Evidently historians who discern
Mycenaean tribal organizations incorporated into the Dorian states of Crete can hardly posit a centuries-long gap
between the expulsion of the Mycenaeans (by a native revolt) and the Dorian seizure of Crete.
This question of substrate elements will arise later in the discussion of the Doric dialects of the Peloponnese. They
are, however, of relevance to the nature and status of the 'Myce-naean' presented in the Linear B tablets. It will
appear that the Doric dialects of Corinth and Argos contain datives in
that are an important characteristic of
the Aeolic dialects and that this evidence is strengthened by ancient tradition that Corinth was an Aeolic settlement
before the coming of the Dorians. If, then, on such evidence we may posit the presence of a dialect exhibiting an
important innovation of Aeolic in the northern Peloponnese at the end of the Bronze Age, this would be additional
reason for believing that the 'Mycenaean' of the Linear B tablets is a standard chancellery language which masks
the contemporary spoken dialects. It is noteworthy that
< previous page page_55 next page >
< previous page page_56 next page >
Page 56
datives occur in the Argolis, for it is in Mycenae that we should expect apriori the chancellery language to have
originated. That there was such a single origin has been underlined by the recent finds of tablets at Thebes. The
Linear B scholar finds himself in a familiar world with the same scribal conventions and administrative practices.
The administrative language was learned by the scribes as part of their general training. This makes it unlikely that
any future accretion of Linear B tablets will enable the linguist to break out of this magic circle. For our
knowledge of the dialects of the Mycenaean Age we shall still have to rely on inferences suggested by a
comparative study of the later dialects, to which we turn in the next chapter.
< previous page page_56 next page >
< previous page page_57 next page >
Page 57
III
The Greek Dialects
1. Introductory
The Mycenaean civilization came to a violent end during the twelfth century B.C. The Linear B script was the
instrument of a centralized palace-based economy, and there is no evidence for its use outside the narrow circle of
palace scribes. Thus the script was bound up with the activity of the palace as the brain and nerve centre of a
vulnerable, over-organized society. Consequently when the palace was destroyed, society fell apart and the script
ceased to have a function. After the collapse there ensued an illiterate 'Dark Age', greatly reduced in population,
wealth and resources, that lasted something like four centuries. When literacy dawns again as the gift of the
Phoenicians,1 the linguistic picture of Greek is totally different. Instead of the rigid uniformity of the Mycenaean
'Kanzleisprache', a supra-regional administrative language, we find a congeries of local dialects, some
characteristics of which may be gathered from the inscriptional evidence. These developments were the result of
historical forces and events which must largely be deduced from the linguistic evidence itself. While political
disunity is matched by an extreme dialect fragmentation, each centre having its own individual dialect, nevertheless
the plethora of patois evidently falls in the first instance into four distinct groups2 with a geographical distribution
which is particularly instructive in the primary 'colonial' area of Asia Minor and the offshore islands (see Fig. 4 and
pp. 64-82 for the historical implications and the differerent colonial periods).
I. Arcado-Cypriot. The dialect of Cyprus has an unmistakable affinity to that of Arcadia in the central highlands of
the
1 On the Greek alphabet, see Chapter VII.
2 The gaps in our knowledge must be stressed. We have few early inscriptions and, by the time that they
become numerous, they exhibit the influence of the 'Common Language' (the
, see below). The fact
remains that we can list clusters of diagnostic features which suffice for the establishment of family groups.
< previous page page_57 next page >
< previous page page_58 next page >
Page 58
Peloponnese and of Triphylia, the southern canton of Elis. The relation of this group to Mycenaean will be
discussed below, but the significant geographical fact is that Arcadian in the alphabetic period is completely
surrounded by speakers of Doric (see next group).
II. West Greek. This comprises two sub-groups (a) Doric and (b) North-west Greek. The Doric area in Asia Minor
is restricted to the coastal region facing the southern Sporades, which extend from Rhodes to Cos and Calymna. A
series of island stepping-stonesCarpathos, Crete and Cytheraleads back to the mainland source of this colonial
area, the south and east Peloponnese comprising Messenia, Laconia, Corinth and thence eastwards to Megara,
where it impinges on Attic territory. The main concentration of the North-west Greek dialects is in Aeolia, Locris
and Phocis (Delphi), while Elis offers a bridge-dialect between NWG and Doric.
III. Proceeding northward in Asia Minor we next enter Ionic territory, which extends from Halicarnassus as far
north as Smyrna.1 This area, including the off-shore islands (the northern Sporades and Chios), is the region of
East Ionic. To the west we have the Ionic of the Cyclades, with West Ionic in Euboea and Oropus (in north-east
Attica). Closely related to Ionic is Attic, essentially the language of Athens, and the whole group is known as
Attic-Ionic.
IV. The coastal area north of Ionia, stretching from Smyrna to the Hellespont, was occupied by speakers of a
dialect known as 'Lesbian', since our main sources for the dialect come from the offshore island of Lesbos. This
dialect shares important characteristics with the dialects of Thessaly and Boeotia and the whole group is known as
'Aeolic'.2 The two mainland representatives are strongly infused with West Greek features (see below).
1 According to Herodotus (1. 150) Smyrna was wrested from the Aeolians by exiled Ionians from
Colophon. On the Aeolic features of the dialect of Chios, see p. 87. Again according to Herodotus, Doric
had once been spoken in his native city of Halicarnassus.
2 Linguists use the term in this purely technical sense. For the ancients 'Aeolic' meant the dialect of Sappho
and Alcaeus (see below). It was also used in a much wider sense to designate what was neither Attic-Ionic nor
Doric. For this 'residual' group comprising Aeolic and Arcado-Cypriot the term 'Achaean' is widely used today
(see below).
< previous page page_58 next page >