Introduction: Gordion and its environs in the Hellenistic period
Scientific research at Gordion, the capital of the Phrygians, has demonstrated that the site was also a significant center in the region during the Hellenistic period[1] . The Hellenistic settlement is focused on the Gordion Citadel Mound, an area of 8-10 hectares[2] . During the 1993-1997 excavation seasons, the excavations on the northwestern part of the Citadel Mound overlooking the Sakarya River revealed the construction of new residential buildings and a monumental building at the beginning of the Hellenistic phase, approximately at the time when the Galatians reached Gordion[3] . Apart from tile fragments, few artifacts were found within the floor deposit of the monumental building, whereas a painted bowl fragment dates the final occupation of the building to the beginning of the second century BCE[4] . The destruction, noticeable throughout most of the settlement on the Citadel Mound, has been attributed to the Roman army that reached Gordion in 189 BCE, led by Manlius Vulso[5] . Megarian bowls, black slip pottery, and red slip thin-walled pottery dating to the first half of the third century BCE, recovered during the extensive excavations that started in the 1950s, provide vital information about the Hellenistic period of the region[6] .
When Celtic tribes migrated through the Balkans in the early third century BCE, the kings of Bithynia and Pontus invited some of these tribes to pass through Anatolia as mercenaries, and these eventually settled in Central Anatolia. The Galatians, as these Celts were known, entered Anatolia around 275 BCE and reached Gordion, probably shortly afterward[7] . The Phrygians predominantly inhabited the area. A small number of people of Hellenic origin had migrated to the cities of Gordium (Gordion) and Ancyra (Ankyra, modern Ankara) following Alexander the Great’s conquest of the Persian Empire (334-323 BCE). Whether the Galatians themselves chose to settle in these remote and often arid regions or were allocated these lands by their allies (or even enemies) is still debated. Even after the so-called Elephant Victory of the Seleucid king Antiochus I, the Galatians appear to have been the dominant power in Central Anatolia. Although they were defeated by Attalus I of Pergamon in 230 BCE and by the Roman General Manlius Vulso in 189 BCE, they never fell under Pergamon’s control, although their sphere of action was clearly reduced during that period[8] .
Upon the arrival of the Galatians in Central Anatolia, where Gordion is located, three centers became prominent. After the Hellenistic period, Ancyra, Tavium, and Pessinus became the major Galatian settlements in Anatolia. Ancient sources reveal that the Galatian tribe of the Tectosages settled in Ancyra, the Trocmii in Tavium, and the Tolistobogii in Pessinus[9] . According to Stephanus of Byzantium, using Apollonius of Aphrodisias, who lived in the third century BCE, as a source, the Galatians founded these three cities[10]. Researchers working on Galatia emphasize that these cities should be interpreted as fortresses.
In the literature, the term Galatian refers to the Celtic communities who settled in Anatolia. The Galatians, directed to Central Anatolia by the efforts of the Hellenistic kings, lost their former strength and gradually assimilated following the invasion of the Roman army commanded by Manlius Vulso into Western Anatolia in 189 BCE. Due to the limited number of excavations and surveys on Galatian settlements, it is impossible to know the characteristics of a Galatian settlement. Nevertheless, the remains and finds from the Hellenistic period from this region indicate that Hellenistic culture heavily influenced the Galatians. When Central Anatolia became the Galatian province of the Roman Empire in 25 BCE, the Galatians fell completely under Roman rule[11].
It is suggested that Pessinus, a religious center in the Hellenistic period, became the emporium of Western Galatia in the Roman period and that Gordion lost its regional control and was reduced to a small settlement along the route between Pessinus and Ankyra[12]. However, based on the excavations carried out at the site, A. Goldman indicates that Gordion was not such a small settlement in the Roman period as has been formerly proposed[13].
Gordion’s environs included many settlements in small fortresses, particularly following the arrival of the Galatians in the region. Research carried out in the early 2000s by Vardar and his colleagues in and around Ankara has provided vital information for understanding Gordion and its immediate surroundings during the Hellenistic and Roman periods.
1.5 km southwest of Basri village, a fortress on a hill overlooks the city of Gordion, the Sakarya River valley, and the route of the ancient road to the west. The architectural remains of this fortress are no longer visible. The foot of the hill yielded coarse ceramics from the Byzantine period. Twenty to thirty meters below the southwestern skirts of the fortress, a tumulus is located on a hill 200-250 meters away. Pottery sherds that can be dated to the fifth/sixth century BCE were identified near the looters’ pits on the tumulus[14].
Another Hellenistic settlement near Gordion lies 1.5 km south of Çanakçı village. The buildings on the hill in this area, constructed by cutting and shaving the bedrock, probably date to the Hellenistic period. A wall here built of rectangular blocks with smooth lines and sharp corners probably dates to the Roman period. Pottery found on the surface indicates these two periods as well. The authors state that the area may have been a sanctuary. Ancient quarries were identified 15 m below the steep southern slope of the citadel. Most of the pottery sherds from the dirt road between the village and the citadel date to the Roman period[15].
A fortress settlement is located 3.5 km northwest of Şeyhali village, where the Şeyhali- Polatlı road turns to the Polatlı-Yunak road. The dense and diverse pottery assemblage at the site dates from the Late Phrygian period onward, including the Roman period. Besides monochrome gray plates from the earlier periods, plates with red paint decoration are also noteworthy[16].
Located 1.5 km southwest of Kargalı village there is a settlement that may also have been a fortress. This settlement is rich in Hellenistic and Roman sherds[17].
Another Galatian fortress settlement is located 5 km northeast of Oğuzlar village. Although the precise dimensions of the fortress are currently unclear, the site has a roughly rectangular plan. Large, quadrangular cut blocks were laid at the foundation level. Based on the pottery evidence, the structure probably dates to the Roman period at the earliest. Late Chalcolithic pottery was found on a terrace about 750 meters north of the fortress[18].
Approximately 1 km northeast of Hacıtuğrul village there is a series of three hills. A settlement was identified on and around the largest of these hills, which is 30 m high. The evidence includes a number of building remains of uncertain plan. The pottery assemblage collected from the site dates to the Phrygian, Hellenistic, Roman, Late Antique, and Byzantine periods[19].
In the vicinity of the tomb at Hacıtuğrul, a re-used andesite capital was discovered on a 15-20 m high hill just east of the fountain, and some pottery sherds from the Hellenistic and Byzantine periods were also found on the surface. Some fragments of Greek inscriptions, located 30-40 m south of the tomb, must belong to grave stelae. A lion statue with a broken and missing head was also discovered in the immediate vicinity[20].
New Phrygian inscription from Gordion: Edition of the text
§1. The inscription is reported to have been discovered in spring 2021 by a local farmworker, Mustafa Sivri, while he was tractor ploughing a field (Kıranharmanı Çiftliği Mevkii) immediately adjacent to the modern Yassıhöyük-Beylikköprü road, northwest of Gordion’s Citadel Mound across the Sakarya River. Although the precise findspot is now uncertain, Mr. Sivri recalled and demonstrated to us that it was in the eastern part of the field, i.e. close to the road and thus approximately 250 m from the Citadel Mound (cf. fig. 1). The inscribed stone was then taken to the house of his employer, Şinasi Genç, the landowner living in the nearby village of Kıranharmanı, and kept in the house courtyard. The inscription’s existence was subsequently reported to Mine Çifçi of the Museum of Anatolian Civilizations in Ankara. In June 2022 the museum director Yusuf Kıraç and vice director Umut Alagöz organized transfer of the stone to the museum where it obtained inventory number 35-1-22[21]. The present author was able to study the inscription in the depot of the museum on July 26.
The inscription is cut into a roughly rectangular massive stone slab (cf. figs. 2-4). The slab’s upper, lower, and left-hand sides are relatively regular, albeit rather roughly worked, and appear to be the original faces. By contrast, the extant righthand side of the slab is evidently broken, its condition indicating that the damage probably occurred in antiquity. The missing part of the slab has not been located. The inscription is in general well-preserved, but there is a long albeit rather shallow gouge running diagonally across the inscribed face, presumably made by the tractor plowshare at the moment of discovery. The gouge runs from the second through the sixth line of text, obscuring one or two letters in each line. The right-hand part of the inscription has been lost with the missing end of the slab.
The height of the stone is 43.5 cm, its width in the widest preserved part is 68 cm, and it is about 14-17 cm thick. This means that we are dealing with a horizontally orientated monument rather than a stele, an unusual and interesting feature. The average height of the letters in the first line is 24-25 mm, as also in those other parts of the text which are written more carefully. In some other lines, however, the letters are somewhat smaller: for instance, they are on average only 20mm high in line 6 and 22 mm in line 7. The inscription is the twelfth text on stone found in the Gordion area and accordingly can be attributed the siglum G-12[22].
Text.
1. … ] Antiyokoy ∙ Śilẹṿkoỵ-kẹỵ vac.
2. … ]n manaṇ mlalin śit-t vac. ꓘ
3. … ]ṃetebaẹs eḳ ṃroteś-key vac.
4. … ] Parśaparnas˙ eś-k ṃrey veiṣ́-t Perkạmạṇeiṣ
5. … i]ḅey Gordiyoy puprayọỵ veban vac.
6. … ]ṇin-key oḷvomun ∙ opoś-key iḅẹy vac.
7. … ]ọy ạẹỵ-t maneis vac.
Epigraphical commentary.
§2. Alphabet and ductus.
The alphabet variety found in the inscription is the first – but by no means the last – surprising feature of the text. No other Phrygian text discovered so far in the region of Gordion shows comparable characteristics. Although several details of the ductus are quite unique, the alphabet in general finds parallels in the inscriptions from a different region, namely from the northwestern part of Anatolia where the borderlands of Phrygia merge with those of Mysia and Bithynia. The most exact correspondences are found in two inscriptions of the ‘northwestern’ group: the long and elaborate inscription from Vezirhan (B-05) and the inscription from Üyücek (B-04), which is much shorter and poorly preserved (cf. Brixhe 2004: 42-67 and 32-42 respectively, with further refs.). There are five remarkable epigraphical features which link these two with the new inscription from Gordion:
1. Letter y. The letter y is one of the important diagnostic features that distinguishes the ‘Middle-Phrygian’ alphabet used in the ‘northwestern’ group of Phrygian inscriptions (B-04-07 and graffiti), dateable to the late 6th to 4th century BC, from the ‘Common’ Old-Phrygian alphabet found in the monuments of central (sigla G-), eastern (the inscription from Kerkenes), western (sigla W-), but in part also northwestern (B-01 and B-03) Phrygia, which belong in general to earlier periods. While in the ‘Common’ Old-Phrygian alphabet the letter y represents a vertical hasta with two short side-strokes added to both ends of the hasta on different sides, the shape of the northwestern y is both more complex and more variable (for an overview of the shapes see Brixhe 2004: 30). In particular, in B-04 and B-05 the letter consists of a diagonal stroke set at about 45º, and to the left of it an ‘angle’ set at approximately 90º, one end of which touches the diagonal stroke roughly in the middle. In the new Gordion inscription, the general shape of the letter y closely corresponds to that of Vezirhan and Üyücek, featuring, however, two slight modifications: the diagonal stroke is set here more horizontally (at about 30º-10º to the base line) and the angular element has a smoother outline, thus more resembling a ‘hook’; also the transition between the ‘hook’ and the diagonal element is in general smoother. There can be little doubt that the latter shape represents a more cursive form of the letter resulting from writing on soft materials, such as parchment.
2. Letter s. The letter s represents the next important diagnostic feature of the ‘northwestern’ Phrygian alphabet. In the ‘Common’ Old-Phrygian alphabet, s represents a vertical zigzag line consisting in the earlier periods of multiple (as many as 7-9) short bars, and in the later periods of 3 or 4 longer bars, which in general compares well with the developmental stages of the Greek sigma. In contrast, the ‘northwestern’ s rather resembles m (both the Phrygian and the Archaic Greek), consisting of a vertical hasta with a roughly horizontal threebar zigzag line added to its upper part. In contrast with m, the zigzag part of s is somewhat less pronounced (i.e. it is ‘shallower’), but the distinction is not always easy to discern. In addition, the s of the Gordion inscription has a longer vertical hasta whose lower part descends below the base line (cf. especially the final s in Parśaparnas in line 4 and in maneis in line 7). As far as can be seen, this feature is absent from both the Vezirhan and the Üyücek inscription, although it is difficult to be absolutely sure given the low quality of the published photos.
3. Letter ś. This letter, most clearly visible at the beginning of Śilẹṿkoỵ (first line) and in the middle of Parśaparnas (fourth line), has not hitherto been documented in any other inscription of the Phrygian corpus (but see below). However, its phonetic value is unambiguously suggested by two indications. On the one hand, the shape of the letter is clearly reminiscent of s, the only difference being that the lower end of its vertical hasta curves left instead of continuing below the base line[23]. On the other hand, the appearance of the letter in Śilẹṿkoỵ, which clearly renders the name of the king Seleucus (cf. below), confirms that we are dealing with a sibilant. Accordingly, the letter can be transliterated as ś.
The presence in the Phrygian alphabet of two letters for sibilants is not something entirely surprising. The use of two different letters for sibilants was earlier assumed for Üyücek[24] and Vezirhan[25], although the shapes of ś (allegedly) found there are not identical. Two different letters for sibilants are further found in graffito B-108 from Daskyleion (Brixhe op. cit. 2004, p. 92-93). Strictly speaking, however, none of these texts features the letter ś in a context which provides unambiguous clues about its exact phonetic value. Consequently, Brixhe (op. cit. 2004, p. 26- 28) attempted a different interpretation of the letter, assuming that it may be a regional variant of the arrow-shaped letter usually thought to render a sort of affricate (/ts/ or /dz/); this interpretation was adopted in Obrador-Cursach op. cit.: p. 33-38. In a recent article,[26] I already put forward arguments against this interpretation, suggesting that the central Phrygian alphabet in fact also knows two different letters for sibilants, thus mirroring the situation assumed earlier for the northwestern Phrygian alphabet. The second letter probably rendered all the different sorts of ‘unusual’ sibilant, from palatal ç to alveolar ʃ to the sharp (geminate) ss. In contrast, the arrow-shaped letter rendering the affricate ts/dz should be kept apart. The evidence of the new Gordion inscription now once again confirms this interpretation.
Τhe distribution of the letter in the inscription agrees with the phonetic interpretation suggested in Oreshko op. cit. 2022[27]. In Śilevkos and śit-t the letter appears before I and thus stands in all probability for a palatalized variant of s (ç). In two further cases, eś-k and ṃroteś-key, it appears after e, which might imply that we are dealing with a similar case of palatalization. However, in ṃetebaẹs, perkamạṇeis and maneis one finds the usual s, while in opoś-key the letter ś appears after o. All this seems to imply that it is rather the following k which triggered a change of articulation, in these cases probably the shift to the alveolar ʃ. The same probably applies to veiṣ́-t, if the letter here is indeed ś (and not s, cf. below). For Parśaparnas one may assume that the triggering factor was the preceding r. It is noteworthy that a parallel to this change is found in Lycian, where one finds the spelling Parzza (TL 44c: 2 and 14) for the very same ethnic name on which Parśaparnas is likely based (cf. below). This use of ś is contrasted with the use of simple s at the word ends in ]ṃetebaẹs, Parśaparnas, perkamạṇeis and maneis.
Moreover, the present clear attestation of ś unexpectedly sheds a new light on the Üyücek inscription. A close examination of the photo given in Cox-Cameron 1932 (cf. fig. 5) reveals that a letter of identical shape is in fact also present in this inscription. It is found in the last but one (in conventional numeration 6’) line of the inscription and was mistaken for n by Cox and Cameron (1932: 38-39) and left without any interpretation by Brixhe (2004: 37, cf. his drawing on p. 34). Besides providing an important link between the Üyücek and the Gordion inscription, the discovery is also important in another respect. It disproves the existence in the Üyücek inscription of an ś having the shape of the Lydian letter š. In fact, none of its three alleged attestations in the inscriptions is seen in any way clearly and its reading was in part already doubted by Brixhe (2004: 34-39). In all probability, in all these cases we are dealing with t.
4. Letter u. In both its occurrences in the inscription (in puprayọỵ and olvomun), the letter has an ‘inverted’ orientation, with the diagonal stroke added to the ‘back’ of the letter (i.e. the right side in the sinistroverse orientation). Such forms of the letter are sporadically found in the inscriptions from central Phrygia (cf. M-01c, W-10, G-204 or G-245), but here they are associated exclusively with graffiti executed in a somewhat sloppy style. In contrast, both Vezirhan and Üyücek feature ‘inverted’ u as the standard form of the letter.
5. Letter e. The shape of the letter is quite particular in the inscription: it consists of an ‘angle’ with a more or less smooth upper part – sometimes approaching the shape of a ‘hook’ – and two side-strokes added to it on the left; it is somewhat smaller than the inscription’s average letter size. Again, the letter has a very similar shape in the Vezirhan and Üyücek inscriptions, though here there is usually a somewhat longer ‘tail’ underneath.
In addition to these specific features linking the three inscriptions, one may note the closely comparable shapes of a, which is centrally symmetrical,[28] d, which is, in contrast, asymmetrical featuring a somewhat shorter right and a somewhat longer left hasta, the very small o, and r with a rather small loop. The shape of k in the Gordion inscription is somewhat different from that found in Vezirhan-Üyücek where it consists of a vertical hasta with a semicircular stroke to the left of it that does not touch the hasta. However, the same principle is present in the Gordion inscription: both k in Śilẹṿkoỵ-kẹỵ and that in -key at the end of line 3 consist of a vertical hasta and an ‘angle’ which does not touch it.
§3. Arrangement of the text.
The text is sinistroverse (i.e. written from right to left) except for one separate letter k which appears close to the left edge of the stone on the level between the second and the third lines (cf. figs. 2-4). On the whole, a sinistroverse text orientation is not unusual for Phrygian monuments since these do not give special preference to either direction of writing. It is noteworthy, however, that all the relatively late monuments of the ‘northwestern’ group (B-01 and B-04-07) are exclusively sinistroverse.
The lines terminate at different distances from the left edge. The reason for this irregularity is apparently the wish to not divide one word between two lines: when it seemed impossible to accommodate one full word in the available space, it was preferred to leave the line end uninscribed. Moreover, the space left at the end of line 6 is so large that one wonders if the next line should begin a new syntactic unit rather than simply a new word. Similar irregular line ends are also found in the Vezirhan inscription (Üyücek is too broken to be sure about this), but in other comparable Phrygian inscriptions on stone blocks (W-11 or B-07) there is continuous writing, with words being divided between two lines when necessary.
This concern to start the new line with a full word gave rise in line 4 to a curious phenomenon. Apparently having miscalculated the available space, the scribe had to transfer the final part of the last word onto the previous line in the space immediately above (perkamạṇeis). As far as one can see, this phenomenon is not attested in any other Phrygian text. It is known, however, from several Lydian inscriptions. Most systematically, the transfer onto the upper line is employed in LW 10 where one finds it in five lines (ll. 8, 11, 15, 18 and 20); besides that it is attested in LW 8 (nã-qiš in line 6), LW 71 (šrmlịš in line 1) and in LW 115 (kopšiš in line 5 and Pajafuliš in line 7)[29]. The presence of this practice in Lydian inscriptions clearly correlates with the strong tendency not to divide words between lines: in the entire Lydian corpus one can identify only four certain cases of continuous writing (LW 6, 8, 9 and 108)[30]. This makes a stark contrast with the Greek writing tradition in which the scribes only very rarely cared about not breaking the words between two lines.
Another curious feature of the text arrangement is the presence of blank spaces which separate words, or to be more precise, prosodic units consisting of a word and a clitic. For the most part, the blank signs are employed regularly and the word-structure of the text is perceptible on the visual level. However, in several cases the blank spaces are missing (or are not large enough to be clearly identified as such). This is especially the case in the second half of line 4, where not only are blank spaces missing but also the letters are set more densely – clearly reflecting a desperate attempt by the scribe to squeeze a long word into the line. In line 6 a blank space is missing before olvomun.
Lastly, one has to note that there are at least two instances when the words are separated by dots rather than simply by blank spaces. This is the case with Antiyokoy and Śilẹṿkoỵ in the first line and with olvomun and opoś-key in line 6. Such a dot is possibly present also after Parśaparnas in line 4, although this is less certain. It is not quite clear if the dot is a separate punctuation sign or is simply an alternative form of word division. In line 6, the dot appears to be separating two syntactical units, which may also apply for line 4 (cf. discussion below), but Antiyokoy˙ Śilẹṿkoỵ-kẹỵ very probably belong to the same syntactic unit. In the latter case, however, there is virtually no blank space between the words, and the dot may have been added later in order to make the division between the words clearer.
§4. Observations on the readings
Line 1. There are no identifiable traces immediately before Antiyokoy, which implies that the word is preceded by a blank space. The identification of the fourth letter in Śilẹṿkoỵ as e raises little doubt since the curve of the ‘hook’ can be seen clearly. The side strokes were apparently incised only slightly. The next letter is damaged, but the phonetic context makes its reading as v quite certain. The last two letters of -key are damaged but the rest are seen clearly enough to exclude other identifications.
Line 2. The ending of mana- is destroyed by the gouge mentioned earlier, with only the lower part of the vertical, slightly inclined hasta in the rightmost position extant (a short stroke to the right of it is apparently a secondary indentation). The identification of the letter as n is suggested first of all by the circumstance that both the preceding and the following word end in -n, implying that we may be dealing with one syntagm in acc.sg. This identification is further supported by the distance between mana- and the first letter of the following word, which corresponds well to the rather broad shape of n used in the inscription. Also, the trace of a diagonal stroke vaguely discernible under the gouge agrees with the identification.
The fourth letter of the next word is ambiguous, as it can be taken either for an l, inclined forward a little more than in other cases (cf. the first l in the word and those in Śilẹṿkoỵ and olvomun), or for a d with a shorter second hasta than that observed in Gordiyoy and having a somewhat ‘rampant’ position. It is noteworthy that the same ambiguity is present in the Vezirhan inscription[31]. The reading of the present letter as l seems preferable. The slightly greater inclination in the direction of writing is not very surprising, and some other letters in the inscription also demonstrate this tendency, cf. especially the first three letters in veban and b in ṃetebaẹs. A tilt in the opposite direction, combined with the shortness of the second diagonal hasta, would be more difficult to explain.
Line 3. At the beginning of the line one can see only a rather long straight diagonal stroke beginning in the upper part of the line. This can correspond to either l or m. Both the length of the stroke and the absence of any traces in the lower part of the line speak for an m. The b of ṃetebaes is unusually big and incised deeper than other letters in the word. It is difficult to explain this oddity. In theory, it might be a conscious attempt to emphasize the letter or the word it contains, although this would be odd given that the letter seems to stand in the middle of the word. Unfortunately, the meaning of the word is quite obscure, thus giving no support to this interpretation. Another possibility would be to see in it the result of a correction, although it is not immediately obvious what would have been the initial variant of the letter.
The second and the third letters of the next word are somewhat damaged by the gouge. The traces of the second letter consist of a vertical hasta with a downward diagonal stroke which starts, importantly, somewhat lower that the top of the vertical hasta which is still clearly discernible despite the gouge. This combination of features can correspond only to k (cf. especially the shape of k in Antiyokoy). The traces of the third letter well correspond to m – with the first inclined hasta still vaguely discernible under the gouge – and the attestation of μροτις in W-11 confirms the reading. The interpretation of eḳ ṃroteś as two words is suggested by comparative evidence (cf. below).
Line 4. In the space before the first clearly seen letter (p) there appears to be a vague horizontal stroke in the lower part of the line, which in theory might correspond to the lower part of an ś. However, although the stone is relatively well preserved in this part, it is impossible to identify any other traces of an ś, or indeed any other letter. Given that the reading Parśaparnas makes good sense as one word, the above-noted horizontal stroke is best interpreted as later damage to the stone in the blank space before the word.
The letters in the second part of the line are somewhat smaller and more densely set, and blank spaces are missing. At the very end of the line the letters are even more crowded. The lower part of the letter following vei- is somewhat damaged, but its identification as ś (rather than s) is more probable since there is no continuation of the vertical hasta downwards. The small letter following pe- is somewhat indistinct but its rounded top and the traces in the middle definitively speak for an r. The letter following k is also somewhat problematic epigraphically, as one can clearly see only a vertical and a long diagonal hasta. However, the fact that the letter appears between kr- and -m- unambiguously defines it as a vowel, so the only sensible reading is a. The last two letters in the line are slightly damaged but can still be identified without much ambiguity. The first is a, its arch and even the horizontal stroke being seen distinctly enough. In the last letter, only the short diagonal stroke can be clearly discerned, and whatever the case, this should be attached to a vertical hasta (which can be figured out only vaguely). Morphological considerations (cf. below) allow the letter to be identified as n rather than l, with the second diagonal stroke at the very edge of the stone broken off. Three letters transferred onto the upper line can be read without much ambiguity as -eis, with the last letter being only slightly damaged. The analysis of the sequence as veiṣ́-t perkạmạṇeiṣ follows from morpho-syntactical considerations (cf. below).
Line 5. The space before the first observable letter of the line (e) is damaged but traces of a semicircular element are still readily discernible. The diameter of the circle is rather large, thus better corresponding to the upper loop of b (cf. esp. that seen in ]ṃetebaẹs) rather than that of r which is unusually small in the inscription. The ending -bey corresponds to that of ibey found in the next line, and the meaning of the word agrees well with the context (cf. below), so the restoration [i]bey appears probable. The gouge affected two last letters of the next word beginning with pupray-. The traces are, however, clear enough and leave little doubt in the reading of the final part as -ọỵ.
Line 6. The gouge damaged two letters of the last word in the line but the remaining traces are clear. The first letter can be identified as b, as no other letter has a similar upward-curling lower part, and the second can be identified as e. The reading is confirmed by the attestation of comparable forms (ibey, ibeya etc.) elsewhere in the Phrygian corpus (cf. below).
Line 7. Due to the damage to the stone, the reading of the second word in the line presents several difficulties. The first letter is unambiguously a, and the fourth letter is t as both its upper horizontal hasta and the lower part of the vertical one can be clearly discerned. In the third letter, one can discern a smooth hook-like upper part and a nearly horizontal stroke in the lower part, indications that speak for a y. The second letter is the most damaged one. Its general shape seems to be quite comparable with the following y but it is clear that it cannot be the same letter. Most probably, it is the other letter which features a hook-like shape in its upper part, e. The reading is corroborated by attestation of the form aey elsewhere.
Philological commentary.
§5. Owing to space limitations, only a brief philological and linguistic commentary essential for understanding the basic contents of the text can be offered here. A more detailed discussion will be presented elsewhere.[32]
Antiyokoy˙ Śilẹṿkoỵ-kẹỵ: the words, connected by the enclitic conjunction -key ‘and’, clearly represent Phrygian forms of the well-known Greek names, Antiochus (Ἀντίοχος) and Seleucus (Σέλευκος), which are strongly associated with the kings of the Seleucid dynasty. As there are no clear indications that the inscription is an official text (a decree or suchlike), the names most probably are part of a dating formula, which agrees well with their appearance in the first line. The ending -oy is probably dative, although its interpretation as genitive cannot be fully excluded.
manaṇ. Formally, the word is probably acc.sg., like the preceding and the following words. No exact counterpart of the word is found elsewhere in the Phrygian corpus. However, etymologically the word can be connected with μανκα ‘stele’ or ‘grave monument’, well attested in the Neo-Phrygian corpus and probably based on the PIE root *men- ‘think, remember’. The term can in general be interpreted as ‘monument’, corresponding semantically to Greek μνημεῖον. A late form of the word, μανη, previously misinterpreted to be a part of the word *κορο[υ]μανη, is probably found in the Nakoleia Monument (2.2 [130], cf. Obrador-Cursach op. cit.: 527-529).
mlalin. The word is a hapax in the Phrygian corpus, and as far as one can see, the consonant cluster ml is not attested in any other Phrygian word. A similar word, mλola, is found, however, twice in a Lydian funerary inscription LW 5 (line 3). The context of the latter text implies that mλola designates a part of a chamber tomb (wãnas). The Lydian word can have either a general meaning, as ‘part’ (cf. Gusmani op. cit. 1964, s.v.) or ‘place’, or be a more specific term for ‘burial chamber’ or another type of burial installation. As i-stems are rather rare in Phrygian, it is unlikely that the stem vowel of mlali- results from a spontaneous change of stem class in the process of borrowing from Lydian to Phrygian. Rather, the Phrygian word can be interpreted as an adjectival form going back to *mlaliyan. As an adjective, mlalin possibly makes a combination with the preceding manaṇ.
śit-t is a form of the demonstrative pronoun si- with an emphatic particle -t. The form probably results from śin-t (attested in the Vezirhan inscription, ll. 1 and 8), with an assimilation nt > tt, and is thus masculine acc.sg. The pronoun may well be connected with the preceding manaṇ mlalin.
]ṃetebaẹs. Given the similarity of the final part of the word with such forms as edaes or estaes, it can be interpreted as a verb in aor.3.sg. It is not quite clear whether metebaes is a full word, or some part of it is missing. No directly comparable form is attested in the Phrygian corpus and the root of the verb cannot at present be identified with certainty.
eḳ ṃroteś. The second part of the letter sequence eḳṃroteś finds a rather exact parallel in μροτις attested in the inscription W-11, 4, suggesting an analysis eḳ ṃroteś. The word ṃroteś, which may be either a nominative singular or nominative plural, can be interpreted either as ‘dead’ (as in Obrador-Cursach op. cit., p. 95 and 305) or as ‘man, mortal’ (probably < *mrotiyas, cf. OP martiya- or Skr. martya-). As for the element eḳ, it is not attested elsewhere in the Phrygian corpus. Although it exactly corresponds to the Greek preposition ἐκ ‘from’, their identity is unlikely in view of both the grammatical case of ṃroteś (ἐκ is used exclusively with genitive in Attic-Ionic) and of its complete absence in other Phrygian inscriptions. Instead, given the presence of the Lydian syntagm eś-k ṃrey further on in the text, one may tentatively suggest that eḳ renders the usual Lydian clause-initial particle ak. This interpretation agrees well with the position of eḳ after a verb (]ṃetebaẹs), which both in Phrygian and in Lydian usually take clause-final position.
Parśaparnas. The word can be identified as a composite personal name of Persian origin. Its second part -parna- clearly corresponds to northwestern Iranian farnah- (= Avestan xv arǝnah-) ‘(divine) splendour’ or ‘glory’. The first part is most probably based on the ethnic name of the Persians, Pārsa, cf. esp. Māda-farnah- ‘Glory of the Medes’33. The name *Pārsa-farnah is not directly attested in other sources but its existence was earlier hypothesized by Rüdiger Schmitt[34] on the basis of Πασιφέρνης.
eś-k ṃrey. Neither eśk nor mrey is attested elsewhere in the Phrygian corpus and the former sounds patently alien in Phrygian. In contrast, both words would make good sense if read in Lydian. The former exactly corresponds to Lydian eš=k, the common demonstrative pronoun ešt (neutr.nom./acc.sg.) followed by the enclitic conjunction =k (with the elision of -t before k). The word mrey in its turn finds a close counterpart in mruj ‘stele’,[35] the common designation of this sort of monument in Lydian. The combination eš=k mruj is directly attested in LW 3: 1, LW 8: 1, and LW 26a: 1. Although it is very surprising to find a Lydian syntagm in the Phrygian text, its identification makes good sense in the context, also agreeing well with other clues, such as the probable presence of yet another Lydian term (mlalin) and the Lydian features in the inscription’s graphic layout (cf. above), as well as the geographic origin of the probable author of the text.
veiṣ́-t Perkạmạṇeiṣ. The first word can be identified as a form of Phrygian possessive pronoun 3.sg. ‘his’, followed by the emphatic particle -t (the same as in śit-t). The second word closely corresponds to Περγαμηνός, an ethnic adjective derived with the common suffix -ην- from the name of Πέργαμον. The correspondence Greek η = Phrygian a < ā is regular (e.g., Phr. matar = Gr. μήτηρ ‘mother’), and the appearance of k instead of γ in Greek does not look surprising in view of other evidence for devoicing of the voiced stops in Phrygian (e.g., Phr. mekas vs. Gr. μέγας ‘big’)[36]. Veiṣ́-t Perkạmạṇeiṣ clearly represent a single syntagm which can be interpreted as ‘his Pergamenes’, the possessive pronoun apparently picking up Parśaparnas of the preceding clause. Grammatically, the ending -eis can be interpreted either as acc.pl. or dat.pl. As the preceding eś-k mrey likely represents acc.sg., the interpretation of veiṣ́-t Perkạmạṇeiṣ as dat.pl. seems more plausible.
Gordiyoy puprayọỵ. The first word is clearly the Phrygian form of the name Γόρδιον. In all probability, it is dative or locative, thus meaning either ‘to Gordion’ or ‘in Gordion’. The word puprayoy finds a close correspondence in p̣upratoy attested in the last line of the Vezirhan inscription (B-05: 13), settling the exact reading of the first letter there. As suggested earlier,[37] the respective passage of Vezirhan likely corresponds to the Greek beneficiary formula found after the Phrygian text, but exact correspondences between the semantic units in the Greek and the Phrygian passages are not fully clear. Now, the combination of puprayoy with Gordiyoy implies that the root pupra- renders the idea of ‘movement’ to a place, and consequently p̣upratoy of the Vezirhan inscription should correspond to Greek ἥκοντι ‘coming’ (part.pres.dat.sg.). The word puprayoy is probably a noun in dat.sg. meaning ‘comer’ or ‘wayfarer’.
veban. As I have argued in an earlier presentation,[38] the word veban does not mean ‘grave’ as it has been previously understood (cf. Obrador-Cursach op. cit., p. 239-240 with further refs.), but rather it has a more or less abstract positive semantic. This is suggested first of all by the passage of the Vezirhan inscription just discussed (l. 13) where the word appears in the apodosis of the beneficiary formula. The present context appears to exactly correspond to that of the Vezirhan inscription. The original meaning of veban may be ‘protection’, but on the synchronic pragmatic level it probably simply expresses a general positive wish, thus roughly corresponding to English ‘good fortune’ or Turkish Hayırlı uğurlu olsun.
olvomun. The word is not found elsewhere in the corpus. However, its root can be tentatively identified in the word olvos possibly attested in the Germanos inscription (B-01: 8), where it was earlier read oyvos (cf. Brixhe-Lejeune op. cit., p. 63-66). Neither of the two contexts provides a clue for its possible meaning. Etymologically, one may consider two possibilities. Firstly, one may compare it with Greek ὀλοός ‘destructive, deadly’ (< *ολοϝός), an adjective based on the verbal root which underlies ὄλλυμι ‘ruin, destroy’, in which case olvomun would make part of a negative formula. The second possibility would be to compare the root of olvomun with Greek ὅλος ‘whole, entire, complete’ which goes back to *holu̯o- < PIE *sol(H)-u̯o-.
opoś (opos). Again, the word is a hapax in the Phrygian corpus. Formally, the word is directly comparable with Latin opus (gen. operis) ‘work, labor, deed, action, accomplishment etc.’ which goes back to *opos-/opes-, and its cognates in Indo-Iranian (cf. Skr. ápas- ‘work’, Av. apah- ‘work’ etc.). Despite the fragmentary context, the presence of the word with such a meaning makes good sense, as it can be taken as a reference to either ‘military deeds’ or to the ‘monument’ as a structure.
iḅẹy. The word is attested in the Vezirhan inscription (line 10), and is probably connected with the form iben attested in line 2 of the same inscription, and with the adjective ibeya found in the Germanos inscription (line 3). Its earlier interpretation as a pronominal form (cf. Obrador-Cursach op. cit: p. 249 with further refs.) appears likely, but the connection with the root of the PIE reflexive pronoun *sVbh e- is unconvincing. Rather, one may compare the word with Lydian epaj and Lycian ebei, both meaning ‘here’. All the forms can be explained as old loc.sg. based on the PIE demonstrative pronoun *(h1 )e-be/i-.
ạẹỵ-t. The word aey is attested in the Areyastis inscription (W-01a) and in the Germanos inscription (line 8), and likely represents a longer form of ae found in one of the inscriptions from Midas City (M-01f). The final -t is probably an emphatic particle. The attestation of the word before maneis precludes its interpretation as a verb (cf. Obrador-Cursach op. cit., p. 162 with further refs.), and at the same time supports its interpretation as a preposition as suggested by Brixhe[39]. The available contexts suggest a meaning ‘to’ or ‘for’.
maneis. The word possibly represents another grammatical form (acc.pl. or dat. pl.) of the noun mana- ‘monument’ attested earlier in the text. Alternatively, it may be another derivative based on the same PIE root *men- ‘think, remember’ and mean ‘memory’. If the latter is correct, the combination ạẹỵ-t maneis may be interpreted as a counterpart of the common Greek collocation μνήμης χάριν (μνείας ἕνεκεν) ‘in memory’.
Translation.
§6. Needless to say, the translation given below is tentative, since without a full context and in the absence of direct parallels in the Phrygian corpus the interpretation of some words is no more than a learned guess. I have avoided, however, using a question mark with every uncertain word, in order to not to encumber the text. The Phrygian text is rendered in English, and the Lydian and supposedly ‘Lydianizing’ words are given in italicized French. It should be kept in mind that a significant portion of the text – possibly as much as 50% – is missing at the beginning of each line.
1. …] (under) Antiochus and Seleucus.
2. …] monument funeraire this
3. …] he …-ed. And the mortals/dead
4. …] Parsaparnas. Et cette stèle for his Pergamenes
5. …] to the wayfarer [he]re to Gordion – good fortune!
6. …] and … perdition/safety. And the work here
7. …] … to the monuments/in the memory.
Geographical and ethnocultural background of the text.
§7. A number of indications suggest that the inscription, although found at Gordion, originates from a different ethnolinguistic and cultural background. These indications are: the alphabet, linguistic features, and the two names found in the text. The combination of these factors allows the origin of the tradition reflected in the inscription to be pinpointed with fair accuracy.
As noted above, the inscription’s alphabet demonstrates similarities with the northwestern group of Phrygian monuments and especially with the Vezirhan and Üyücek texts. Given that the inscription is the first one from Gordion that can be called ‘Middle-Phrygian’ – and is moreover later than those from Vezirhan and Üyücek – one may ask if the similarity of script cannot reflect simply a sort of Middle-Phrygian alphabetical koine which spread, with minor variations, throughout the entire Phrygian dialectal zone in the 4th–3rd centuries BC. This perception is refuted by a number of facts. Firstly, among the stone inscriptions found in the Gordion region there is one clearly late text. This is G-11, consisting of three short sequences on three sides of a trapezoidal stone block (Brixhe op. cit. 2002, p. 31-37) which can be identified as a field boundary marker[40]. The letters are much more rounded than those usually found in the graffiti from Gordion, and the shape of s – a smooth three-bar zigzag line – is clearly late. The stone most probably dates to the 4th or even the early 3rd century BC. Secondly, the letter shapes found in the numerous graffiti from the Gordion citadel show no significant similarities with those used in the present inscription. Although this is due in part to the fact that many of these graffiti date to earlier periods, some of them should belong to the 4th and early 3rd century BC, despite many uncertainties concerning the chronology of individual specimens. In contrast, the alphabet of the graffiti from Daskyleion (cf. Brixhe op. cit. 2004, p. 85-93) demonstrates essentially the same characteristics as found in the monumental inscriptions. Lastly, there is a late inscription, W-11, from the region of Dokimeion (northeast of modern Afyon) which can be dated to the 4th-early 3rd century (Brixhe op. cit. 2004, p. 7-26). The letter shapes and the style of this text are very close if not identical to those found in contemporary Greek monuments, showing little in common with the Phrygian inscriptions of the northwest. Thus, there is no reason to speak about a late ‘koinization’ of the alphabet in Phrygia: as far can be seen, each sub-region retained its own writing tradition until all of them were obliterated in the process of Hellenization during the 3rd century BC. The similarities between the alphabet used in the present inscription and that of Vezirhan and Üyücek can only be interpreted as an indication of a northwestern origin for the former. Moreover, a small but highly significant detail – the correspondence in the shape of the letter ś found in the present text with that found in the Üyücek inscription – suggests a link with the more southerly parts of the general northwestern area, i.e. basically Mysia, rather than Bithynia or the region around Daskyleion. The correspondence in the visual organization of the present inscription with that of the Lydian texts (i.e. separate writing of words, and transferring words onto preceding lines) once again corroborates this.
The linguistic features of the inscription are no less telling. The dialect used is characterized by the same phenomenon of secondary syllable-closing glide -y at the word ends (cf. -key and aey) that one finds in the northwestern zone and in W-01. But the differences within the group are no less significant. Firstly, the assimilated form śit-t of the Gordion inscription is in contrast to the sin-t found in the Vezirhan inscription. Secondly, the Gordion text shows Lydian influence. Although the presence of mlalis may simply be due to the genre specifics of the inscription, the presence of an entire syntagm eś-k mrey strongly implies that the inscription’s tradition derives from a region where Lydian was spoken along with Phrygian. This further confirms the evidence of the alphabet.
Lastly, there are two names, of which the ethnic Perkamaneis (‘Pergamenes’) is especially important. Although the exact context of the mention of the Pergamenes remains unclear, the very fact of their appearance is highly significant precisely against the background of the probable western links of the text. The region of Pergamon corresponds amazingly well with what the evidence of the alphabet and language implies. Pergamon is located approximately 250 km to the west of the Üyücek findspot, and geographically both places belong to Mysia. On the other hand, the region of Pergamon is immediately connected with Lydia via the valleys of the Caicus and the Lykos, and the temple of Athena in Pergamon is the findspot of a Lydian inscription (LW 40). There can be little doubt that a Lydian community was present in pre-Hellenistic Pergamon, and this circumstance presents a ready explanation for the use of the Lydian syntagm and the possible Lydian funerary term in the Gordion inscription. The concurrence between the three strands of evidence – epigraphy, language, and the ethnic name – is so good that one can with confidence conclude that the origin of the Gordion inscription is connected with the Pergamon region, despite the absence of direct epigraphical parallels there.
§8. The fact that the new inscription G-12 mentions a person bearing a Persian name introduces a curious nuance into the picture. It is not fully certain that Parśaparnas was the author of the text, though this is probable. In any event, the use of the pronoun ‘his’ with ‘Pergamenes’ establishes a close link between the two and strongly implies that Parśaparnas originated from the same geographic and probably linguistic background. Again, this is not very surprising, since relevant evidence bearing on the longstanding Persian settlement in the Pergamon region is provided by Xenophon.
At the end of his Anabasis (7.8.8-23) he describes a military undertaking against a Persian named Ἀσιδάτης who lived with his family and household in what appears to have been a sort of ‘fortified villa’ (τύρσις , lit. ‘tower’) located in the Caicus plain to the east of Pergamon, probably not far from the river Karkasos (mod. Ilias Çay). This τύρσις was not the only landed property of Asidates, because anticipating a second attack by the Greeks he escaped to the ‘fortified villages below the town of Parthenion’ (possibly northeast of Pergamon), which apparently also belonged to him. Moreover, Asidates was by no means the only Iranian present in the surrounding area. Xenophon (7.8.15) reports that, alarmed by shouting and beacon fires, quite a number of other troops came to his aid: a certain Ἰταμένης,[41] probably another Persian noble possessing lands in the region, with his own force; a large contingent from a place called Κομανία[42] which consisted of 80 Assyrian hoplites and Hyrcanian horsemen, who are called ‘mercenaries of the [Persian] King’ (βασιλέως μισθοφόροι), as well as 800 light troops (πελτασταί); and an unspecified number of troops from Parthenion, Apollonia (located in the eastern part of the plain) and other nearby villages. It is doubtful that all the light troops consisted of ethnic Iranians – in all likelihood, the bulk of them were local men under the command of Persian officers, cf. the employment of Mysian troops as ‘outposts’ (προφυλακή) by Pharnabazus, mentioned by Xenophon (Hellenica 4.1.24). However, the households of Asidates and Itamenes consisted most probably of Persians and the same may apply to at least some of the men from the other villages. Thus, it is clear that already around 399 BC the population of the eastern part of the Caicus plain had a non-negligible Iranian component, being in this respect similar to the plains of Lydia[43].
It would be tempting to see in Parśaparnas a later representative of the Persian nobility settled in the Pergamene region. Like Asidates or Itamenes, he may have been the commander of a military contingent recruited for the most part from the local inhabitants of the surrounding area and which at some point was sent to Gordion (for the historical context cf. below). It is noteworthy that in this role Parśaparnas represents a close parallel to Pharnabazus, son of Artabazos, a Persian officer probably stemming from the dynasty of satraps of Hellespontine Phrygia, who, according to Plutarch (Eumenes 7.1), served in the army of Eumenes of Kardia around 322 BC[44]. In all likelyhood, Parśaparnas was born not in Persia but in Mysia, representing the third if not the fourth generation of the Persians living there, and despite his Iranian name his (second) mother tongue was probably the Mysian dialect of Phrygian.
Nature of the associated monument.
§9. Due to the fragmentary character of the text, the genre of the inscription and the nature of the monument with which it was once associated are not immediately evident. Still, available clues suggest that it was a funerary monument, though probably a rather unusual one.
The first indication comes from the findspot of the stone. As noted, it was found some 250 m to the northwest of the Gordion Citadel Mound, across the modern Sakarya River, which corresponds to the territory of the so-called ‘Outer Town’. In the Hellenistic period both the ‘Lower Town’, located closer to the Citadel Mound, and the ‘Outer Town’ were largely abandoned, and at least some parts of their territory were used for burials[45]. It is noteworthy that a kilometer to the west of the inscription’s findspot a Hellenistic tomb was found under a tumulus (Tumulus O)[46]. Given this situation, the simplest – though not the only possible – interpretation of the stone would be a funerary inscription. At the same time, its interpretation as an official text is virtually excluded.
The second clue is provided by the shape of the stone. As mentioned, it is not a stele but a horizontal slab. This necessarily means that the stone was a part of some larger structure. This circumstance can be brought into agreement with the indications of the text, subtle though they are. If manas in line 2 does not necessarily imply that we are dealing with a complex funerary structure – in the Nakolea Monument μανη seems to designate the stone pillar itself – two other words are more suggestive. Mlalis is probably connected with the Lydian mλola which refers to a part of a tomb chamber, most probably a separate ‘room’. The form maneis in line 7, if interpreted correctly as pl. of mana, would also confirm that the monument consisted of several elements. A further corroboration of the funerary character of the text is given by the word ṃrotes which is etymologically connected with ‘death’, although in theory it can mean ‘mortal’.
However, is not fully clear to whom the burial monument once belonged. On the one hand, there is Parśaparnas, and it is possible that he, as the only person mentioned by name in the extant part of the text, is the owner of the monument and the author of the inscription. On the other hand, there is the beginning of the phrase ‘Et cette stèle for his Pergamenes …’. Although it remains obscure how mrey relates to manan mlalin, the clause gives the impression that the monument was somehow connected also with the Pergamenes, who probably were, as suggested above, the men serving under the command of Parśaparnas. In fact, taking into consideration the particular circumstances of the presence of a Pergamene contingent at Gordion, and the likely personal bond between Parśaparnas and his men, it is entirely possible that the structure was conceived both as a funerary monument for the former and as a monument commemorating the death of the latter – but not necessarily serving as their grave.
Whatever the case, it is unlikely that the structure was simply an underground tomb comparable with that discovered beneath Tumulus O. The text makes it reasonably clear that it was supposed to be read. Especially eloquent in this respect is line 5 which seems to explicitly address those coming to Gordion. This circumstance is crucial for the reconstruction of the shape and location of the monument: the inscription was very probably set up in a way that it could easily be noticed by those who passed by. This implies first of all that the monument stood close to the road which approached Gordion from the west – which is of course not very surprising given the examples from other necropoles. Second, the inscription should have been in a visually prominent position. Possibly the simplest option would be to assume that the inscribed block served as a lintel above the entrance to a free-standing funerary construction. In contrast, it is very difficult to imagine how a horizontal block could have been combined with the kind of tumulus familiar at Gordion, as these typically lack dromoi.
Furthermore, in light of the Mysian links of the inscription and the owner(s) of the monument, there is no particular reason to expect that it would be a typical Gordion tumulus. Rather, it should be a type of funerary construction specific to the region where Parśaparnas and his soldiers came from. Free-standing funerary monuments were, however, not typical of the local tradition in central western Anatolia, in the way they were in Lycia for instance. And yet, they are not completely unknown, though belonging to a different cultural stratum: that associated with the Persians. The most famous example of such a monument is Taş Kule, a rock-cut free-standing tomb to the east of Phacaea, located only about 70 km from Pergamon. This monument finds a close parallel in the tomb of Cyrus at Pasargadae but also integrates some local architectural elements, so it was constructed in all probability by a Persian noble living in the region[47]. Another example of a ‘Persianizing’ monument, although constructed in a very different style, is the Pyramid Tomb at Sardis belonging to the early Achaemenid period. As Parśaparnas very probably hailed from the same class of Persian or ‘Persianizing’ elite as the builder of Taş Kule – which he very probably had a chance to see – it is not impossible that his monument constructed at Gordion had a comparable appearance. Unlike Taş Kule, however, it would have been constructed from stone blocks, which in all probability were subsequently robbed to be re-used as building material for some later, more prosaic construction.
Dating and historical context.
§10. The dating ‘(under) Antiochus and Seleucus’ may seem at first glance to be imprecise, as the majority of the kings of the Seleucid dynasty bore one of these two names – in particular all the kings who ruled between 306 BC and 162 BC (when Demetrios Soter took the throne). In reality, however, this indication establishes a rather precise dating, far superior to anything obtainable for the vast majority of other Phrygian inscriptions.
The crucial indication comes from the sequence of the names. The sequence Antiyokoy Śileṿkoy-key straightforwardly implies that the former was the older and principal co-regent and the latter the junior one. In theory, this combination would suit several pairs of Antiochus-Seleucus, but in practice there is only one sensible possibility from a historical point of view, that of Antiochus I Soter (281- 261) and his elder son Seleucus. The latter is indeed known to have been the official co-regent of his father from 280 BC until his (suspected) conspiracy and death in 268 BC. As co-regents they are found also in Greek inscriptions from Asia Minor, frequently in the dating formula βασιλευόντων Ἀντιόχου καὶ Σελεύκου[48]. In contrast, Antiochus II Theos (261-246 BC) named his eldest son Seleucus (II Callinicus) as his successor only shortly before his death in 246 BC. Antiochus III Megas (222-187 BC), on the other hand, co-ruled with his son Seleucus IV Philopator only for two years, in 189-187 BC, which is too short to be credible as a dating in an inscription (and anyway it is too late).
The dating of the inscription to the co-reign of Antiochus I Soter and his son in 280-268 BC perfectly agrees, first of all, with two general historical considerations. Firstly, the very use in the text of an epichoric Phrygian alphabet and the Phrygian language strongly speaks for a relatively earlier date within the Hellenistic period. All other known Middle-Phrygian inscriptions are older, belonging generally to the time span between the late 6th and the late 4th century BC; the latest hitherto known inscription is probably W-11 belonging to around 300 BC. As far as one can judge, the use of epichoric alphabets and languages had been gradually ousted by the use of Greek as the official language following the Macedonian conquest of Anatolia. This was the case not only in Phrygia but also in Lydia, Caria and Lycia, where there are only a few inscriptions that can be dated after ca. 320 BC. The process of linguistic and alphabetical Hellenization was probably not linear, and it is quite possible that in some regions the ancient alphabetical tradition was retained for longer than was the case in the principal political centers such as Sardis and Xanthos. But it seems quite unlikely that there remained many such places after two generations from the conquest (i.e. after ca. 270 BC).
Second, there is the Galatian factor. The Galatians invaded Asia Minor in the winter months of 278/77 BC via the Hellespont and appear to have been active first in the northwestern parts of Anatolia, chiefly Bithynia and western Paphlagonia, undertaking raids in different directions, and aiming especially at the rich cities of western Anatolia[49]. The chronology of the Galatian expansion in Central Anatolia is not very clear, but it is generally agreed that the Galatians – more precisely probably the tribe of the Tolistobogii – occupied the region around Gordion during the 260s BC[50]. Obviously, this led to major changes in the ethnic composition and the culture of the region. The construction of a monument connected with a Persian noble and the Pergamenes, as well as the very use of the Phrygian alphabet and language at Gordion after ca. 260 BC, seems to be quite improbable.
§30. What is especially important, however, is that the dating to the early part of Antiochus I’s reign presents a sensible historical framework for the appearance of a Pergamene contingent led by a Persian at Gordion. As a result of the Battle of Corupedium in Lydia in 281 BC, in which Seleucus I Nicator defeated Lysimachus, the greater part of Anatolia (excluding only Cappadocia and Pontus) came under the control of the Seleucid dynasty. At this point, the region of Gordion also certainly belonged to the Empire of Seleucus, if it had not been taken already after the Battle of Ipsus in 301 BC. A year earlier (282 BC), Philetaerus, the commander of the fortress of Pergamon and the keeper of its treasury, defected from Lysimachus to Seleucus[51]. Philetaerus is also known to have purchased the body of Seleucus, who had been killed only several months after the battle of Corupedium by Ptolemy Ceraunus. After having the body cremated, Philetaerus sent the ashes to Antiochus. These actions probably established a good personal relationship with the latter, securing the position of Philetaerus as the ruler of Pergamon and giving him a degree of independence under Seleucid suzerainty. Using the significant funds he had at his disposal, Philetaerus not only engaged in building activities at Pergamon but also tried to forge good relationships with other polities in the region, by supporting them with money or troops. An especially relevant piece of evidence is found in an inscription from Cyzicus (IMT Kyz Kapu Dağ 1485) which, among other things, mentions on two occasions the dispatch of Pergamene troops for ‘guarding the territory’ of the city (Il. 6-7: φυλακὴν τῆς χώρας ἴππους πεντήκοντα and l. 13: φυλακὴν τῆς χώρας). The second occasion, which can be dated to 278/77 BC, is clearly the invasion of the Galatians, which is explicitly mentioned later in the text (l. 18-19: ἐν τῶι πολέμωι τῶι πρὸς τοὺς Γαλάτας γ[ενομένωι]).
It appears highly likely that the presence of the Pergamene contingent under the command of Parśaprnas at Gordion should likewise be seen in the context of the Galatian invasion. The evidence concerning Antiochus’ engagement with the Galatians is very scarce, although there can be little doubt that they presented a major challenge to his rule in Anatolia. Ancient sources essentially reduce Antiochus’ encounter with the Galatians to a single event, dubbed in modern scholarship the ‘Elephant Battle’, in which Antiochus defeated the Galatians and thereby curbed their further expansion to the south[52]. Neither the place nor the exact date of the battle -if there was indeed only one battle- is known with certainty, but a date between 276 and 274 BC appears most probable for at least the first major encounter. One may suggest that the Pergamene troops led by Parśaprnas formed one of the military contingents engaged by Antiochus to counter the Galatian attacks. General considerations imply that the battle should have taken place not too far from the mountainous parts of northwestern Anatolia, where the Galatians were operating in the early phase of their migration into Anatolia. The region around Gordion, which quite probably still retained its importance as a fortress in the early Hellenistic period, would have been well-suited as a mustering place for the troops of Antiochus, a number of which he apparently brought with him from the eastern provinces, while others were recruited from the armies of local Anatolian governors such as Philetaerus of Pergamon. It seems thus quite probable that Parśaprnas and his men saw the battle elephants of Antiochus trampling the mustachioed Galatian warriors somewhere to the northwest of Gordion, which introduces one more colorful touch into the extraordinary story told by the inscription[53]. Whether Parśaparnas was killed in the battle or not, the monument should have been constructed at Gordion when it was still felt to be a relatively safe place. These observations allow the dating of the inscription to be narrowed to 276-270 BC. The story sheds yet a little more light on the history of early Hellenistic Gordion, showing that even in this period it remained an interesting place which saw far more curious things than Alexander cutting his knot.
FIGURES
