Dialectal Differences at Knossos

[This text, “Woodard 1986,” was originally published as an article in Kadmos 25 (1986), 49–74. In this online version, the original page-numbers of Woodard 1986 will be indicated within braces (“{“ and “}”). For example, “{51 | 52}” indicates where p. 51 of the original article ends and p. 52 begins. These indications will be useful to readers who need to look up references made in previous scholarship to the original printed version of this article.]

In his article “Les différences dialectales dans le mycénien,” [1] Ernst Risch has argued that documents from Pylos preserve traces of two distinct dialects of Mycenaean Greek. On the basis of frequency of occurrence, Risch calls these “mycénien normal” and “mycénien spécial.” These two dialects are distinguished by the following three features:

The occurrence of Special Mycenaean features is limited to the works of a small number of scribal hands. However, as explicitly stated by Gregory Nagy, [3] it is never the case that some scribal hand X exclusively uses Normal Mycenaean forms while some other hand Y uses only Special forms; instead, those few hands which exhibit Special Mycenaean characteristics do so infrequently and sporadically, while usually employing Normal Mycenaean forms. Apparently, those scribes utilizing Special Mycenaean forms were native speakers of a ‘nonstandard’ dialect who most often employed the ‘standard’ dialect but who occasionally lapsed into the use of their native speech. {49|50}

A fourth Special Mycenaean dialect feature has been suggested by Nagy. Mycenaean, like the other South Greek dialects, regularly assibilates *ti to si. Consequently, nominal stems ending in –to/ā– form derived adjectives in –sio/ā-:

There are, however, a few cases at Pylos in which an unassibilated ti alternates with the expected si:

Two of the Pylian scribal hands writing ka-pa-ti-ja are also seen to exhibit one of the Special Mycenaean characteristics identified by Risch: namely, dialect feature 3. Both hands 1 and 41 write e-pa-sa-na-ti for the woman’s name i-pa-sa-na-ti (though 41 ‘corrects’ this spelling by erasing the initial e- and inserting i). On the basis of this distribution, Nagy tentatively proposes that unassibilated t before i is another Special Mycenaean characteristic. [6]

In stark contrast to Nagy’s suggestion, Risch has recently claimed that unassibilated t is not a Special Mycenaean characteristic. Risch’s argument is presented in response to Chadwick’s proposal that the Dorians were already present in the “Mycenaean-speaking area of Greece” in the late second millennium B.C. There, according to Chadwick, they existed as a large middle and lower class populace subject to the rule of the Mycenaean aristocracy. Furthermore, Chadwick goes on to suggest, Special Mycenaean is proto-Doric. [7] However, after examining the distribution of several forms with unassibilated t, Risch concludes that

The remainder of this paper will focus upon the corpus of Mycenaean materials from Knossos. An attempt will be made to identify all examples of the three Special Mycenaean dialect features proposed by Risch as well as all instances of unassibilated t before i. Furthermore, by examining the distribution of these four features, an attempt will be made to determine, first, if certain of the scribes of Knossos can be identified as speakers of the Special Mycenaean dialect, and second, if unassibilated t before i is, in fact, a fourth dialect characteristic of Special Mycenaean.

Three examples of this dialect feature come from the F series of tablets: documents recording divine offerings and rations.

1.1 te-ra-po-ti. Consider first of all tablet F 193 (written by scribal hand “124” [9] ):

On the basis of form and context, te-ra-po-ti, which designates the recipient of the specified quantity of barley, is most likely to be either nominative singular or dative singular ending in -i. In order to determine which of the two is more plausible, related documents must, of course, be examined. The only other barley tablet by a “124” hand is F 51 (hand “124”d):

This tablet probably contains at least one dative, di-we (= Διϝεί), in a context which is quite similar to that in which te-ra-po-ti is found. Furthermore, tablets in this series which record the distribution of barley and which were written by scribal hands other than “124″(d) [10] offer {51|52} two apparent datives ending in -e—a-*65-ma-na-ke (Fs 3, perhaps the name of a deity) and, more securely, pa-de (probably also a divine name):

Compare pa-de-i (Ga 953) with a fuller spelling of the dative ending. This same set of tablets also provides two examples of what appear to be accusatives of place names with the allative ending -de, i.e. forms which indicate the destination of the barley offerings: o-ja-de (Fs 9) and sa-na-to-de (Fs 2). The nine remaining names of recipients and destinations found on these tablets end in -o or -a and, hence, are grammatically quite ambiguous. However, since none of these names must be interpreted as nominative and since several forms do appear to be in an oblique case (either dative or accusative), it is reasonable to suggest that te-ra-po-ti is, in fact, dative singular.

1.2 *56-ti. The Fp tablets, records of ritual offerings of olive oil, provide the second example of a dative ending -i:

The tablets in this series, [11] most of which are introduced by the name of a month, specify no fewer than ten offering recipients. At least four [12] different recipients are listed a total of fourteen times in the dative case: pa-si-te-o-i (= πάνσι θεοῖhι [13] , ten times), di-we (Fp 1), pa-de (Fp 1, 48) and a-re (“to Ares”? [14] , Fp 14). Four other recipients have names which terminate in -o/a: qe-ra-si-ja (apparently a divine name, [15] seven times), a-ne-mo-i-je-re-ja (= Ἀνέμων ἱερείᾳ, [16] Fp 1, 13 (twice in the latter)), {52|53} pi-pi-tu-na (apparently a divine name, Fp 13, cf. Δίκτυννα [17] ) and si-ja-ma-to (probably a man’s name, [18] Fp 48). Although the case of these forms is obscured by the writing system, they appear in the same contexts as the above-mentioned datives, e.g.

and, consequently, should in all likelihood be interpreted as datives themselves.

In addition to the above-named recipients, the Fp tablets list eight destinations. Of these, seven appear in the accusative case with the allative suffix: da-da-re-jo-de (Fp 1), *47-da-de (Fp 1), di-ka-ta-de (Fp 7), *47-ku-to-de (Fp 13), au-ri-mo-de (Fp 13), a-ka-ta-ra-te-so-de (Fp 5504) and a-mi-ni-so-de (Fp 14, 48). In contrast to the last-named form, Fp 1 lists simply a-mi-ni-so. Notice, however, that its context is identical to that of the two occurrences of a-mi-ni-so-de:

a-mi-ni-so may be a locatival dative used to qualify the noun which it precedes. Compare a somewhat similar use of place names in the locatival dative and locatival instrumental cases at Pylos:

The toponym u-ta-no, which is the only place name never occurring in the allative accusative, is similarly used to qualify a recipient,

{53|54} and, accordingly, may also be locatival dative. However, the possibility that a-mi-ni-so and u-ta-no are both nominatives can certainly not be eliminated.

In contrast to these various names of recipients and toponyms which have been read as datives and accusatives, there is one form which does, in fact, appear to be nominative: namely, the divine name e-ri-nu (= Ἐρινύς) in Fp.1.8:

However, Furumark [20] and Palmer [21] both interpret this form as a dative Ἐρινύι, cf. e-ri-nu-we, KN V 52.

The one remaining name of a recipient, *56-ti, is, again on the basis of form and context, most probably either nominative or dative singular. It has been suggested above that four of the ten recipients listed in this series of tablets can be identified as datives with reasonable certainty; another four can be interpreted as datives on the basis of parallel occurrence; seven place names can be identified as allative accusatives and two others can be tentatively identified as locatival datives. On one occasion, however, the name of a recipient appears to be nominative in form. Given the preponderance of oblique case forms used to identify recipients and destinations, it is, again, reasonable to suggest that *56-ti is dative rather than nominative. This interpretation is strengthened by the occurrence of the dative pa-si-te-o-i in the same line (see Fp 15, hand 138, provided above).

1.3 *56-i-ti. A third example of this dialect feature, one which perhaps represents a variant spelling of the preceding example, [22] comes from tablet Fh 1057:

The 123 Fh tablets, many of which are fragmentary or only a single line in length, record allotments of oil. [23] Among the fifty-two [24] {54|55} occurrences of personal names and occupational terms which are found in these tablets, there are seven dative plurals (e.g. de-ma-si = δέρμασι, Fh 353), one dative singular following the preposition o-pi (du-ru-po, a man’s name, Fh 368), six dative singulars ending in -e for which nominative forms are found elsewhere (e.g. ka-pa-ri-jo-ne, Fh 344, cf. ka-pa-ri-jo, a man’s name, passim) and five forms which, given their context, are apparently third declension dative singulars but for which corresponding nominatives are not found (e.g. do-re-we, a man’s name, Fh 342). [25] The remaining thirty-three [26] forms end in –o/a and, thus, their grammatical case is less certain; moreover, ten of these appear on tablets which are so fragmentary that the context of each form is obscure. The other twenty-three forms, however, contextually parallel the datives discussed above, and, consequently, many of these are probably to be interpreted as datives as well.

The tablets of the Fh series offer a total of twenty forms which have been identified as place names. Fourteen appear in the allative accusative while six terminate in -o/a. Three members of the latter group occur on severely fragmented tablets and two are followed by the term a-pu-do-si (= ἀπόδοσις, ‘payment’). According to Ventris and Chadwick, [27] a-pu-do-si is used to designate quantities of oil which have been sent to the palace from outlying areas; consequently, these two place names could conceivably be either nominative or ablatival genitive.

The same arguments which have been offered for interpreting te-ra- po-ti and *56-ti as datives are, of course, applicable to *56-i-ti. Although evidence for the last-named may be somewhat less compelling, it is still the case that a large number of recipients and destinations are clearly specified by using the dative and accusative cases rather than the nominative; moreover, there are no names of recipients or destinations which must be interpreted as nominatives. Thus, there seems to be sufficient {55|56} motivation for assigning *56-i-ti to the dative rather than the nominative case. [28]

1.4 to-ni. The final example of this feature occurs on tablet V 145:

As is the case with all members of the V series, there are no ideograms on this tablet; however, it appears that the document is concerned with the collection of some commodity.

The first line of the tablet is obscure but the next three lines appear to begin with titles of officials: u-wo-qe-ne and we-re-we. Both of these titles are also found on C 902 (a cattle inventory tablet from Knossos) along with the frequently occurring title ko-re-te, although in that document the first-mentioned term is spelled u-wo-qe-we; for this form, Palmer suggests *ὐϝοq^w ῆϝες (‘overseers’). [29] On the basis of this pattern, one would expect a-ke-to-ro, the first word in line 5, to be an official title as well—perhaps *ἀρχετροι, cf. ἀρχέτᾱς (‘leader, prince’).

The second form in each of the first three complete lines can be identified as a place name. As pointed out by Palmer, u-du-ru-wo is certainly to be equated with the toponym o-du-ru-we (locatival dative) which occurs on C 902 in association with the above-mentioned u-wo-qe-we:

Furthermore, another tablet in the same series, Co 906, contains the form ka-ta-ra-i: [30] the locatival dative equivalent of the locatival instrumental ka-ta-ra-pi, cf. the similar alternation between pa-ki-ja-si (PY Cn 608) {56|57} and pa-ki-ja-pi (PY Jn 829 and Ma 221). [31] ku-pa-sa is not found elsewhere but its interpretation as a place name is supported both by its context and by the ethnic ]ḳu-p̣ạ-si-ja (V 1043), though the latter piece of evidence must be considered somewhat tentative in the light of the uncertainty of the reading.

Given that u-du-ru-wo, ku-pa-sa and ka-ta-ra-pi all appear to be place names, to-ni can reasonably be interpreted as a place name as well, since it is used in a parallel fashion. Moreover, since both u-du-ru-wo and ka-ta-ra-pi clearly stand in an oblique case (genitive singular and instrumental plural respectively) qualifying the form which precedes, one would expect that ku-pa-sa and to-ni exist in a similar syntacticosemantic relationship with the forms which precede them. If to-ni is in an oblique case, then it must certainly represent another example of the dative ending -i. [32]

Supporting evidence for the interpretation of to-ni as a place name comes from the L series of tablets from Knossos. These documents, along with the closely related Lc tablets, appear to be records of textile goods sent to the palace from various localities. [33] Both ethnics ending in -ja [34] and place names are quite common among the tablets of these two groups. On tablet L 192 is found a form which can be interpreted as an ethnic adjective built upon the place name of which to-ni is the dative:

Usually these place names and ethnics are written in large characters and introduce the tablets on which they are found; however, on L 469 and Lc 7377, place names (Phaistos and Kudonia respectively) are written in a context very similar to that of to-ni-ja:

{57|58} The first line is perhaps to be read as me-ki-ta (‘largest’); po-ku-ta designates a class of men. [35]

The meaning of ]ẉẹ-ri-jo-jo is obscure. [36]

It is true, of course, that the possibility that to-ni-ja is a non-ethnic adjective modifying the cloth ideogram TELA² cannot be eliminated. Such adjectives are not unknown on these tablets: e.g. pu-ka-ta-ri-ja (perhaps ‘of double thickness,’ L471, 474), ta-ra-si-ja (‘amount issued for use,’ Lc 535, 536). [37]

Finally, it is somewhat interesting that tablet L 192 is written by scribal hand “124,” the same hand which is responsible for V 145 (the tablet on which to-ni is found). Of the seventy-seven tablets in the L series which are products of identified scribal hands, only four are the work of hand “124” (L 104, 178, 192 and 5599). None of the fifty-four Lc tablets is by “124”. [38]

In contrast to te-ra-po-ti, *56-ti, *56-i-ti and to-ni, the great majority of third declension datives found in the documents from Knossos terminate in -e. In Table 1, a summary of the occurrence of these forms is presented. {58|59}

2.1 pe-ma. The only example of this Special Mycenaean characteristic which is found at both Pylos and Knossos is pe-ma (= σπέρμα). This spelling is used by scribal hand 135 on three different occasions (Ga 674, 675, 680) as well as by the unidentified hand of tablet Ε 1569. The Normal Mycenaean pe-mo never occurs at Knossos, however.

2.2 ka-ra-ma-to and ke-ma-ta. Two examples of this feature are found on tablet V 684:

In line 1, e-re-pa-to apparently represents the genitive singular ἐλέφαντος (‘of ivory’). Given the parallel occurrence of ka-ra-ma-to and ke-ma-ta, {59|60} both Palmer [39] and Ventris and Chadwick [40] suggest as possible interpretations κλασμάτων and κέρματα respectively (both meaning ‘fragments’). However, for the former, one would have expected the nominative plural *ka-ra-ma-ta; for this reason, Palmer [41] questions the correctness of the spelling of the term as it appears. The one remaining form, ka-so, is perhaps, like ἐλέφαντος, the name of some material; [42] hence, V 684 can be cautiously translated as

2.3 de-ma-si. The Fh series of tablets provides a fourth possible example:

In the first edition of Documents, [43] Ventris and Chadwick interpreted de-ma-si as δέρμασι, meaning ‘in leathern bottles,’ cf. Odyssey 2:291. However, as pointed out above, in these documents allative accusatives are usually used to specify destinations (e.g. ra-ma-na-de) and datives to specify recipients; hence, Palmer [44] suggests that de-ma-si does not designate the container in which the oil was sent but some unidentified recipients: a suggestion with which Chadwick [45] now concurs.

There is, however, one other tablet on which de-ma-si is found, though, unfortunately, the tablet is broken:

On the obverse side of the tablet, ]pte-si is probably to be restored as dative plural ra-]pte-si, cf. Fh 1056: {60|61}

ra-pte, which occurs frequently at Pylos, has been etymologized by both Palmer [46] and Ventris and Chadwick [47] as *ῥαπτηρ (‘sewing-man,’ cf. ῥάπτης). Furthermore, Ventris and Chadwick [48] point out that the use of the adjectival derivative ra-pte-ri-ja to describe a-ni-ja (= Homeric ἡνία, ‘reins,’ PY Sb 1315.2) suggests that the *ῥαπτῆρες were sewers of leather or saddlers. Accordingly, de-ma-si on Fh 5432, as well as on Fh 353, may indeed represent δέρμασι and be translated as ‘for the leather (pl.)’ or ‘for the leather goods.’ Certainly saddlers would have need of oil in the practice of their craft. Similarly, tablets Fh 5428 and Fh 5435 (?) record the allocation of oil to a recipient whom Ventris and Chadwick [49] tentatively identify as a ‘tanner,’ i.e. *ϝρινήϝει: [50]

It must be pointed out, however, that although this interpretation of de-ma-si appears attractive, the fragmented condition of tablet Fh 5432 must be borne in mind. [51]

The documents from Knossos also provide several examples of the Normal Mycenaean counterpart to this feature, i.e. ο from IE syllabic nasals. These are presented in Table 2. {61|62}

Of the four dialect features under consideration here, this one is evidenced by the fewest examples. Alongside the man’s name a₃-ki-wa-to, found on the land tenure tablet Uf 987,

one finds a₃-ke-wa-to on two tablets by hand 117, both of which record contributions of sheep to the palace:

Similarly, alternating with the man’s name qa-mi-si-jo,

{62|63} one finds qa-me-si-jo in a long list of men’s names by hand 101:

4.1 ku-ta-ti-jo. An example of this feature is provided by the ethnic adjective formed from the frequently occurring toponym ku-ta-to (ku-ta-i-to on C 902). On three occasions, scribal hand 136 spells the ethnic ku-ta-ti-jo (twice on tablet Ga 419 and once on Ga 673). The same form is used by the unidentified hands of tablets Wb 5662, X 7897 and G 820. [53] In contrast, ku-ta-si-jo is found on tablets Dv 1237 and 1394 (hand 117), where it appears to be used as the name of a shepherd, cf. a ₃ -ku-pi-ti-jo (= Αἰγύπτιος), a shepherd’s name on Db 1105; also pu-na-si-jo, the ethnic of pu-na-so which appears in a list of men’s names (B 806). Furthermore, tablet X 7891 (hand 118?) offers the form ku-ta-i-si- (cf. ku-ta-i-to above). [54]

4.2 ru-ki-ti-jo. Λύκτος, a place name known from Homer (Iliad 2:647), occurs frequently at Knossos, where it is consistently spelled ru-ki-to according to Ventris and Chadwick. [55] The ethnic adjective ru-ki-ti-jo appears in works by scribal hands 103 (E 668, 670, Ln 1568; on the last-named tablet, the ethnic is feminine and appears to be used as a woman’s name, cf. the discussion above), “124” (Xd 168), “124”? (Xd 314), 136 (E 749, Ga 415) and 201 (C 902) as well as on tablets Og 833 and X 37, both written by unidentified hands. An alternant form which shows assibilation is not found. [56]

The syllabic spelling ki-to for a cluster kt is, of course, quite unusual. The possibility that ru-ki-to is actually not the spelling of Λύκτος but of some otherwise unattested toponym *Λυκιστος cannot be ruled out {63|64} (cf. Homeric Λύκαστος, also Iliad 2:647; Palmer [57] rejects the identification with Λύκτος and indicates that Λύκαστος is geographically more likely but still problematic orthographically). In this case, the ethnic would be ru-ki-ti-jo (i.e. *Λυκιστιος) even in an assibilating dialect since assibilation of t before i is blocked by a preceding s. Vilborg [58] points out that other irregularites are known to exist in the representation of kt clusters, e.g. wa-na-ka-te-ro for *ϝανάκτερος (Knossos, Pylos and Thebes), perhaps under the influence of wa-na-ka (= ϝάναξ). Ventris and Chadwick [59] suggest that the spelling of the place name could have been modified under the influence of the ethnic, or it “could indicate an obscure vowel in the non-Greek form of the name.” Still more obviously, one could suggest that the non-Greek form was simply *Lukitos and that the middle vowel was syncopated at some time after the fourteenth century. Notice that assimilation ultimately produced Λύττος. [60]

4.3 ti-ri-ti-jo. As a final example, consider the ethnic adjective ti-ri- ti-jo formed from the place name ti-ri-to. Ventris and Chadwick, [61] noting the similarity which this form bears to Τρίτα (Hesychius records: Τρίτ(τ)α· οὕτως ἡ Κνωσσὸς ὠνομάζετο), tentatively suggest the reading *Τριτος. Both Hart [62] and Melena [63] point to a possible link between the place name ti-ri-to and the Triton river. If either (or both) of these associations is correct, then ti-ri-ti-jo also shows an unassibilated ti sequence. This ethnic is found on tablets Ε 749 (hand 136), Og 833 (initial syllable reconstructed but reading probable) and X 1385 (the last two tablets are both products of unidentified hands). [64] {64|65}

These three examples of the fourth dialect feature occur a total of nineteen times. In addition to these, another seventy-four occurrences of forms ending in -ti-jo/ā are found at Knossos. Some of these spellings probably represent words ending in -stios/ā, but others may well come from a nonassibilating dialect; however, there is not sufficient evidence to identify these.

There are, of course, a number of forms in which t has undergone assibilation. These are presented in Table 3.

{65|66} A summary of the examples of Special Mycenaean features found at Knossos in the works of identified scribal hands is presented in Table 4. In this table and in those which follow, dialect feature 4 will be treated as a Special Mycenaean trait for the sake of comparison.

Jean-Pierre Olivier [65] has proposed the following schema for the organization of the scribal bureaucracy at Knossos: {66|67}

• West Zone of the Palace
  • 1. Specialized Bureau I — Concerned with the administration of sheep flocks
  • 2. Specialized Department I — Concerned with the administration of the textile industry (Oliver defines a department as “un ensemble de bureaux.” [66] )
  • 3. Specialized Department II — Concerned with recording amounts of spices, aromatic products, honey and divine offerings
  • 4. Specialized Bureau II — Concerned with recording amounts of oil (Olivier indicates that he is uncertain if this should be considered a separate bureau or should be incorporated into Specialized Department II.)
  • 5. Nonspecialized Collection I — Composed of a diverse group of documents found in the Room of Chariot Tablets (find-place C) and written by the various scribal hands labeled “124” (see below)
• North Zone of the Palace
  • 6. Nonspecialized Collection II — Composed of tablets treating a wide variety of subjects; traces of twenty-seven identified and at least thirteen unidentified scribal hands are found
• East Zone of the Palace
  • 7. Specialized Bureau III — Concerned with the administration of sheep flocks
  • 8. Bureau I — Concerned primarily with the keeping of personnel records (Olivier suggests that there may actually be two distinct bureaus here.)
• Palace Exterior
  • 9. Bureau II — Concerned primarily with the cataloguing of military equipment

In Table 5, these bureaucratic units are listed along with the scribal hands which are associated with them. In some cases, a question mark following the number of a scribal hand indicates that the identity of that hand is to some extent uncertain (101?, 102?, 118?, 138?, 141?), while in other cases it indicates that the proper association of a scribal hand {67|68} with a particular bureaucratic unit is questionable (119?, 132?, 133?, 202?, 206?).

Tables 6 and 7 list by bureaucratic unit those scribal hands showing Normal Mycenaean traits and those showing Special Mycenaean traits. Following each scribal hand number, the dialect feature or features used by that hand are indicated in parentheses. In some cases, these features do not actually occur on tablets produced by the particular bureaucratic unit with which the scribal hand is associated. For example, while hand 138 uses Normal Mycenaean feature 1 on three different tablets (a total of four occurrences), none of these tablets comes from Specialized Department II. One of these was recovered among the tablets of Nonspecialized Collection I, where it appears to have been misplaced, and two were found in a clay chest some distance removed from the various find-places associated with Specialized Department II (though, in a note, {68|69} Olivier [67] suggests that the tablets in the clay chest should perhaps be linked with this department). Each of those scribal hands which appeared with a question mark in Table 5 (for one of the two reasons mentioned above) is marked with an asterisk in Tables 6 and 7. A question mark in Tables 6 and 7 indicates that the identity of a scribal hand is to some extent uncertain (thus, 103 (1, 2?, 4) means that hand 103 uses features 1 and 4, and a hand identified as 103? uses feature 2).

An examination of Table 6 discloses that as many as twenty-two of the hands belonging to the scribal bureaucracy may exhibit Normal Mycenaean characteristics. Tables 1, 2 and 3 and the discussion of dialect feature 3 reveal that these twenty-two hands are responsible for ninety-two examples of Normal Mycenaean features. However, it is {69|70} only prudent to take into account the various uncertainties discussed above; consequently, these totals will be adjusted in the following way:

22 hands	92 examples
a. -4	-4	hands whose identity is marked as uncertain in Table 6 (112?, 113?, 114?, 223?)
b. -0	-3	hands whose identity is some cases marked as uncertain but in other cases marked as secure in Table 6 (103?, 118?, 128?)
c. -0	-4	hands whose identity is in some cases marked as uncertain but in other cases marked as secure in Table 1 (102?, “124”?, 141?)
d. -0	-1	hands whose identity is in some cases marked as uncertain but in other cases marked as secure in Table 2 (131?)
e. -1	-4	hands whose bureaucratic affiliation Olivier questions by reason of uncertain identity (138?, see Table 5) [68]
f. -1	-1	other hands whose bureaucratic affiliation Olivier questions (132?, see Table 5)
———	———
16 hands	75 examples

Thus, at least sixteen scribal hands affiliated with the scribal bureaucracy offer a total of seventy-five examples of Normal Mycenaean dialect features. [69] {70|71}

As indicated in Table 4, there are ten different scribal hands in whose works Special Mycenaean dialect features have been identified. With the exception of hand 138, each of these is securely associated with some unit of the scribal bureaucracy; hence, at least nine hands affiliated with the bureaucracy exhibit no less than twenty-three examples of Special Mycenaean features (i.e. the number of examples offered by hands whose identity or affiliation is unquestioned). [70]

By utilizing the above tables, one can make several interesting observations about the distribution of these hands:

A) Only three of the hands exhibiting Special Mycenaean traits are found exclusively outside of the West Zone of the palace: 101, 117 and 201 (hand 136 occurs in both the West and North Zones). Hence, the following ratio exists for scribal hands using Special Mycenaean forms: {71|72}

In contrast, the corresponding ratio for hands exhibiting Normal Mycenaean characteristics is much smaller:

A comparison of the two ratios reveals that the relative occurrence of Special Mycenaean forms in the West Zone is almost five times greater than that of Normal Mycenaean forms. [72]

B) Two of the three Special Mycenaean hands which are found exclusively outside of the West Zone (101 and 117) use Special Mycenaean feature 3. This is the only Special Mycenaean trait which they exhibit, and, conversely, they are the only hands exhibiting this trait. Of the various examples of Special Mycenaean features presented in Table 4, qa-me-si-jo and a₃-ke-wa-to are, perhaps, the least satisfying. Notice that if these are not, in fact, Special Mycenaean forms, then the S.M. ratio given above is dramatically increased to 13:1.

C) If the hand responsible for tablet Fh 462 is, in fact, 141 (i.e. the hand marked 141? by Olivier; see Table 7), then all of the West Zone scribal hands showing Special Mycenaean traits belong to Specialized Department II except for “124.” Moreover, recall that Olivier expresses uncertainty concerning the status of Specialized Bureau II (see above). If this unit is, indeed, to be combined with Specialized Department II, then, again, all but one of the West Zone Special Mycenaean hands belong to this department.

Notice that many of the scribal hands in Specialized Department II also exhibit Normal Mycenaean characteristics. Such is not surprising, {72|73} however, since, as was pointed out above, those scribal hands using Special Mycenaean features at Pylos were found also to employ their Normal Mycenaean counterparts. Moreover, while the ratio of total number of Normal Mycenaean forms to total number of Special Mycenaean forms is 75:23 or 3.26:1, the ratio of Normal forms to Special forms used by scribal hands belonging to Specialized Department II is only 15:14 or 1.07:1. Thus, Normal Mycenaean forms are used with relative infrequency by the hands of Specialized Department II.

The one West Zone scribal hand showing Special Mycenaean characteristics which is certainly not associated with Specialized Department II is hand “124.” According to Olivier, [73] “124” actually designates at least ten different scribes. These scribes are distinguished from all others by the physical appearance of the tablets which they produced as well as by their distinct graphic style. That is, in the works of “124,” and nowhere else, a number of unusually formed symbols are found to co-occur. It seems hardly coincidental that a group of scribes regularly using nonstandard syllabograms should also exhibit nonstandard dialect characteristics.

With the exception of hands “124” and perhaps 141, individual scribal hands at Knossos are not found to exhibit more than a single Special Mycenaean trait; however, those scribal hands which use Special Mycenaean dialect features are seen to be concentrated in one area of the palace: namely, in the West Zone and particularly, to a lesser or greater extent (depending upon the status of hands 138 and 141), in Specialized Department II. This pattern of distribution certainly suggests that speakers of the Special Mycenaean dialect were members of those bureaucratic units housed in the West Zone. Notice that this pattern has emerged more clearly in treating unassibilated t before i as a fourth dialect feature of Special Mycenaean; in other words, unassibilated t is found in the works of scribal hands which belong to the same bureaucratic units as those hands using Special Mycenaean features 1 and 2.

The evidence from Knossos obviously supports Nagy’s proposal concerning the Special Mycenaean status of ti rather than that offered by Risch. As indicated above, Nagy points out that Pylian hands 1 and 41 both exhibit unassibilated t (in ka-pa-ti-ja) as well as at least one other Special Mycenaean dialect feature. Risch, however, in his survey of examples {73|74} of unassibilated t at Pylos, overlooks the use of ka-pa-ti-ja by hands 1 and 41. [74] The presence of forms with unassibilated t in the works of hands not previously identified as Special Mycenaean may, in fact, as Nagy [75] himself has suggested, indicate that there were far more speakers of this dialect within the scribal community at Pylos than was earlier realized.

The presence of forms with unassibilated t in Special Mycenaean does not necessarily suggest that this dialect is to be identified as proto-Doric. As Risch [76] points out, it is quite possible that unassibilated forms were found in the South in an early period. Thus, given that Special Mycenaean has none of the other distinctive North Greek characteristics, it may simply represent a South Greek dialect in which assibilation has failed to occur.

In conclusion, I contend that, although some of the individual pieces of evidence presented in this paper must be considered somewhat tentative, the evidence as a whole suggests first, that certain of the scribes of Knossos were speakers of Special Mycenaean; and second, that unassibilated t before i is a fourth dialect feature of Special Mycenaean.