Yet another look at a possible Mycenaean reflex in Homer: phorēnai

By Gregory Nagy | 2023.09.07

§0. This essay is about a Homeric form phorēnai (φορῆναι), which I argue matches a form that is spelled po-re-na- in a Linear B text from Pylos. I argue here that the Homeric form is a reflex of the same form that was spoken in the Mycenaean Greek dialect that we see attested in that Linear B text. My essay here is a new version, with minimal updatings, of an earlier version, dated 2015.03.01 and titled “A second look at a possible Mycenaean reflex in Homer: phorēnai.” That earlier version, published online by the Center for Hellenic Studies, consisted of two parts. Part I was based on a printed article, “A Mycenaean reflex in Homer: phorēnai,” which was originally published in Minos 29-30 (1994–1995) 171-175. Part II was never printed. The new version here, dated 2023.09.07, replaces the earlier version dated 2015.03.01, since both Part I and Part II have now been rewritten. The page-numbers of the original printed version of Part I, dating back to 1994–1995, are indicated within braces (“{” and “}”). For example, “{133|134}” indicates where p. 133 of the printed text ends and p. 134 begins. I should add that the new version of Part I here is formatted differently in two important ways: (1) the paragraphs are now numbered and (2) the notes have been reorganized. The original numbering of the notes in the original version has changed, since most of the bibliographical references have now been integrated into the prose of my argumentation. Wherever the remaining notes in this new edition contain any specific arguments left over from the printed version, I indicate the old page- and footnote-numbers.

§1. This inquiry is about finding survivals of Mycenaean Greek in Homeric diction. I focus on the Homeric infinitive phorēnai (φορῆναι), which I argue belongs to a dialectal group commonly known as Arcado-Cypriote. The form itself is a linguistic innovation, but the innovation is old, very old. I can say this if I succeed in showing that this form φορῆναι is already very old in terms of Homeric diction. And my case would be even stronger if I also succeeded in showing that φορῆναι is attested already in Mycenaean Greek. The evidence of the Linear B texts written in Mycenaean Greek shows a form spelled po-re-na, and this form, in terms of my argument, actually corresponds to the Homeric infinitive phorēnai (φορῆναι).

§2. This infinitive phorēnai (φορῆναι) is attested four times in Homeric poetry (Iliad II 107, VII 149, X 270, Odyssey xvii 224).[1] The internal evidence of the formulaic system underlying Homeric diction shows that the form is old, deeply embedded in that system, and not a newer artificial creation of the Dichtersprache.

§3. In saying what I just said, I am reviving arguments I presented in an early work (Nagy 1972, hereafter cited by way of a second edition dated 2023.08.21). Arguing against the claim of Ernst Risch (1958:92) that φορῆναι is an artificial creation paired with φορήμεναι (as in XV 310), on the model of the athematic type μιγῆναι (ΙΧ 133, etc.) as paired with μιγήμεναι (VI 161, etc.), Ι pointed out (Nagy 1972 [2023.08.21]:64-65) that no other Homeric verbs with present forms ending in -έω have -ῆναι for infinitive. Instead, we see -ήμεναι, as in καλήμεναι (X 125), πενθήμεναι (xviii 174), ποθήμεναι (xi 110), φιλήμεναι (XXII 265). Also, in the case of athematic aorist pairs like δαμήμεναι vs. δαμῆναι, δαήμεναι vs. δαῆναι, μιγήμεναι vs. μιγῆναι, φανήμεναι vs. φανῆναι, etc., “the type in –ῆναι regularly occurs in the archaic slot of line-final position, or in the secondary conversion-slot immediately preceding the trochaic caesura; the type in –ήμεναι, on the other hand, regularly occurs immediately preceding the bucolic diaeresis” (Nagy 1972 [2023.08.21]:64); further, “the latter slot tends to suit a relatively greater proportion of innovated forms” (again, Nagy p. 64).

§4. As for the dialect features of phorēnai (φορῆναι), they are clearly Arcado-Cypriote from the standpoint of reconstructing backward in time from the first to the second millennium BCE. We may compare such forms as Arcadian ἀπειθῆναι and Cypriote ku-me-re-na-i = κυμερῆναι (Nagy 1972 [2023.08.21]:63, with reference to Thumb and Scherer 1959:133, 169). The two basic features are:

§5. Since the displacement of the type φορέω by the type *φόρημι is a linguistic innovation, the attestation of the form φορῆναι in Homeric poetry can be seen as a most precious criterion for establishing the dialectal affinities of the earliest recoverable dialectal phase of Homeric diction as reconstructed backward in time, from the first to the second millennium BCE.[2] I can say this because we have already seen that φορῆναι, as a linguistic innovation that is exclusive to Arcado-Cypriote, must have become part of Homeric diction at a very early stage in the evolution of that diction. In other words, Homeric φορῆναι has to be explained as belonging to the dialect family of Arcado-Cypriote.[3] And, if I can show that Linear B po-re-na- can really be interpreted as phorēnai (φορῆναι), then we have evidence that the standard dialect of Mycenaean Greek was most closely akin to Arcado-Cypriote.

§6. In the part of their jointly-written book that goes back to the first edition, Michael Ventris and John Chadwick (1956:285) had entertained the possibility that phorēnai (φορῆναι) is attested in the component po-re-na- of the Linear B expression do-ra-qe pe-re po-re-na-qe a-ke in Pylos tablet Tn 316, to be interpreted as dōra-k^ʷe pherei phorēnai-k^ʷe agei.[4] We may “translate” thus into classical Greek: δῶρά τε φέρει φορῆναί τε ἄγει. This possibility was rejected by Leonard Palmer (1963:53, 63, 267, especially p. 63), who argued that po-re-na- is a noun and that a-ke is not to be interpreted as ἄγει.

§7. Palmer’s judgment about po-re-na- (but not about a-ke) has in general prevailed. In the “Additional Commentary” of the second edition of the Ventris-Chadwick book (1973:461), Chadwick interprets po-re-na- as an accusative plural designating “the ten persons who are led to the rite.”

§8. My inquiry returns, with modifications, to the possible reading, first mentioned by Ventris and Chadwick, of po-re-na- as phorēnai. A relevant piece of evidence, I suggest, is the syntax of the following Homeric passage:

§9. The hero Diomedes has just won first prize in a chariot race that is featured as the first athletic contest of the Funeral Games honoring the dead Patroklos. Earlier, at Iliad XXIII 263-265, Achilles had determined that the first prize in this contest would be bipartite: to be given away will be a slave woman (263) and a tripod (264). In the text of the consequent narrative as I have just quoted it, the chariot driven by Diomedes is the first to reach the finish line, and we see the victorious hero jumping down from the platform of the chariot (509) and leaving the task of unharnessing the horse team to his companion Sthenelos (513), a hero who elsewhere functions as the chariot driver of Diomedes whenever the two of them together engage in chariot fighting, as in Iliad V. Here in Iliad XXIII, Sthenelos is left with the task of taking hold of the first prize (511), which is bipartite: there is a slave woman to be taken away, agein (512), and there is a tripod to be carried away, pherein (513). [5]

§10. The collocation of dōke (δῶκε) ‘gave’ plus agein (ἄγειν) ‘to take away’ and pherein (φέρειν) ‘to carry away’ in this Homeric passage is comparable with another collocation that we find in four other Homeric passages. This other collocation involves (1) the same word for ‘give’ (δῶκε / δῶκε / δοίης / δῶκ᾿) and (2) the word phorēnai (φορῆναι), already cited, which is a derivative of pherein ‘carry’ and which likewise means ‘to carry’—or ‘to wear’. In §2 above, I have already listed the four Homeric attestations of the form phorēnai, but now I will also list the direct objects indicating what is being ‘carried’—or ‘worn’: a scepter to carry (Iliad II 107), a set of armor to wear (VII 149), a helmet to wear (X 270), and fodder for a herdsman to carry for feeding a herd of goats (Odyssey xvii 224).

§11. Now that I have these Homeric comparanda in place, I am ready to consider the text written in Mycenaean Greek:

§12. Focusing on the expression do–ra–qe pe–re po–re–na–qe a–ke, at lines r2-r3 of the recto and at lines v2, v5, and v8 of the verso, I highlight the following formal correspondences with the Homeric passage that I already quoted:

§13. Of these four correspondences, the first three are straightforward. Only the fourth one involves uncertainties and calls for debate.

§14. That said, I focus here on my reading of do–ra–qe pe–re po–re–na–qe a–ke at lines 2-3 of one side and at reverse lines 2, 5, and 8 of the other side of the tablet Tn 316. I read dōra-k^ʷe pherei phorēnai-k^ʷe agei, the equivalent of which in classical Greek would be δῶρά τε φέρει φορῆναί τε ἄγει. I translate ‘carries [pherei] gifts [dōra] and takes-along [agei] for the carrying [phorēnai] (of the gifts) …’. In terms of this reading, the subject of the verbs pherei ‘carries’ and agei ‘takes-along’ here is impersonal, in line with the prescriptiveness of the ritual instructions. Similarly, I read i-je-to as a prescriptive impersonal statement, translating it as ‘makes sacrifice’. [8]

§15. Although the verbs pherei (pe-re) ‘carries’ and agei (a-ke) ‘takes along’ and hietoi (i-je-to) ‘makes sacrifice’ are all impersonal, deprived of a personal subject, they all nevertheless share what I call an impersonal subject. The verb i-je-to at lines r2 and v1 and v4 and v8 is correlated with the place-name pu-ro that we read at lines r2+3 and v1+2+3 and v4+5+6+7 and v8+9+10+11. This is the name for the city of Pylos, pu-ro, which would be Pulos in the nominative case. As García-Ramón observes, a nominative Pulos at lines r2+3 and v1+2+3 and v4+5+6+7 and v8+9+10+11 could in theory function as the subject of the verb i-je-to at lines r2 and v1 and v4 and v8. [9] And here is where I apply the idea of an impersonal subject.

§16. In support of this idea, I note that the formatting of the word pu-ro as a headline, so to speak, at lines r2+3 and v1+2+3 and v4+5+6+7 and v8+9+10+11, written in larger characters than the rest of the text, could indicate that it functions as the subject not only for i-je-to but also for do-ra-qe pe-re po-re-na-qe a-ke at lines r2+3 and v2 and v5 and v8. As we will see in a moment, an essential piece of evidence in favor of this syntactical interpretation is the expression a-ke-qe wa-tu at line v1.

§17. The ritual procedure of taking gifts to divinities is well known from the evidence of fifth-century Greek: an ideal example is ἄγειν … δῶρα ἐς τὰ ἱρά ‘to take [agein] gifts [dōra] to the sacred precincts [hiera]’ in Herodotus 1.53.1. In the text of the Pylos tablet, however, we see that the ritual of agein or ‘taking’ the offerings to a sacred precinct is subdivided into ‘carrying’ objects, as expressed by way of pherein, and ‘taking’ persons, as expressed by agein. We see a comparable subdivision in the Homeric passage that I cited, Iliad XXIII 512-513, where the prize that is given as a gift consists of a tripod for the recipient ‘to carry away’, pherein (513), plus a slave woman for the recipient ‘to take away’, agein (512).

§18. The fact that the recipients of the gifts are divinities in the text of the Pylos tablet helps us understand the status of the persons who are being taken to these divinities. In this text, the gift of a votive object or objects is optionally supplemented by the gift of a votive person or persons. And this person or these persons must be votive gifts just as the corresponding objects are votive gifts. So, in terms of my interpretation, these persons are slaves who can be given away as consecrated property, just as the objects are being given away as consecrated property. In the case of the persons who are being given as consecrated gifts, the consecration itself is indicated by the fact that the gender of the persons given consistently matches the gender of the divine recipients.

§19. Still to be explained is the expression a-ke-qe wa-tu at line v1 of Pylos tablet Tn 316, which I have translated ‘and the city [wastu] takes-along [agei] (the gifts)’. This way of referring to the idea of offering gifts would be the least specific way of expressing such an idea. In other words, agei ‘takes-along’ would be an unmarked way of referring to the act of offering gifts; by contrast, pherei ‘carries’ would be the marked way. (I use here the terms marked and unmarked along the lines formulated by Jakobson (especially 1957); details in Nagy 1990 Introduction §§12-16.) In the prescriptive formula a-ke-qe wa-tu, I argue, wastu ‘city’ functions as a common noun in apposition to the proper noun Pulos ‘Pylos’. In terms of my argument, wastu ‘the city’ is an impersonal subject of agei ‘takes along (the gifts)’, just as Pulos is the subject of hietoi ‘makes sacrifice’. And the use of the noun wastu (wa-tu) ‘city’ as the impersonal subject of the verb agei ‘takes along’ here at line v1 is parallel to the juridical use of dāmos (da-mo) ‘district’ as the impersonal subject of the verb phāsi ‘says’ at line 5 of the Pylos tablet Ep 704.

§20. Although I have replicated in Part I most of the content I presented in the original printed version of my work in Nagy 1994–1995, I have omitted some additional arguments I made about po-re-na- in footnote 10 on the printed page 173 there. That is because those arguments are now superseded, in my opinion, by further arguments presented in an essay by Roger Woodard (2018.02.04; also 2020.11.03), where he has more to say not only about Linear B po-re-na- but also about two other forms that he proves to be relevant, Linear B po-re-si and po-re-no-. I mostly agree with his argumentation about these three Linear B forms.

Postscript, 2024.04.20

§21. With regard to my interpretation of Mycenaean do–ra–qe pe–re po–re–na–qe a–ke, it could be counterargued that the two instances of the conjunction –qe are used here to correlate the word do-ra, the noun for ‘gifts’, dōra, with the word po-re-na, which, in terms of such a counterargument, would have to be likewise a noun—so, not an infinitival verb that corresponds, in term of my argument, to Homeric phorēnai (φορῆναι). In my work, however, as also in the work of others (documentation by Woodard 2018.02.04, especially at his §5), I have pointed out some serious morphological and semantic difficulties faced by those who propose that po-re-na is a noun, supposedly coordinated via the conjunction –qe with the preceding noun do-ra, as if the first –qe in do-ra-qe were meant as a syntactical coordinate with the second –qe in po-re-na-qe. In what follows, I argue against such a proposal. In my Bibliography below, I include relevant works cited by Woodard (the most important of those works: Palaima 1996–1997 and 1999).

§22. True, a common syntactical construction in later Greek is the pairing of two nouns X and Y by way of the conjunction τε, which is of course the equivalent of Mycenaean –qe. Thus we can expect coordinations, via τε, of “noun X τε” with “noun Y τε.” But I disagree with such an interpretation in the case of do–ra–qe pe–re po–re–na–qe a–ke, as if do-ra-qe and po-re-na-qe were two nouns coordinated by –qe and –qe. In formulating my disagreement, I have benefited from a consultation with Anna Bonifazi (2024.04.13), hereafter “AB,” whom I thank for her valuable observations.

§23. I start with the fact that, in Homeric diction, the initial τε of any syntagma does not necessarily require a syntactic coordination with any subsequent τε. Here I draw on a most relevant observation by AB, who points me to a typical example—this from Iliad I 361:

χειρί τέ μιν κατέρεξεν ἔπος τ’ ἔφατ’ ἔκ τ’ ὀνόμαζε·

With her hand [χειρί]+τε she stroked him, and she spoke out what-she-spoke [ἔπος]+τε, and by-name-she-called-out-to-him [ἐκ…ὀνόμαζε·]+τε.

§24. In this example the first τε follows the noun χειρί but does not coordinate this noun with the noun ἔπος via the second τε, just as the second τε that follows the noun ἔπος does not coordinate this second noun with the verb ἐκ…ὀνόμαζε via the third τε.

§25. Moreover, as AB has pointed out to me, it is relevant that the generalizing “epic τε,” when it follows a relative pronoun, can actually generalize the meaning of the whole clause that is introduced by the pronoun.

§26. As for the four Homeric passages featuring the infinitive phorēnai (φορῆναι), the contexts of which I have surveyed at §10, AB observes that in two of these passages, Iliad VII 149 and X 270, we see the verb dōke(n) / δῶκε(ν) ‘gave’ in collocation with the infinitive φορῆναι / phorēnai—just as we see the noun do-ra / dōra ‘gifts’ in collocation with po-re-na, which I continue to interpret as phorēnai.

§27. Finally, with reference to what I say at §§15–16 about Pylos tablet Tn 316, where I interpret the use of pu-ro ‘Pylos’ here as an impersonal subject referring to the organizer(s) of the ritual program that is being described in the text, I thank AB (again, 2024.04.19) for making a further observation about such such a usage as I interpret it, where ‘Pylos’ is an impersonal subject that metonymically stands for the community of Pylos as an organizational totality. AB points out to me that it would make more sense for the writer of the text to use such an impersonal subject to be followed by a compounded verb meaning ‘carries gifts as well as takes along (people) to carry (the gifts)’. It would make less sense for the writer to use a personal subject here, since the ‘carrying’ is not being done personally by the organizer(s) of the ritual program involving carriers who are actually carrying the sacral gifts.