This book, a revised version of the author’s Habilitation in Indo-European linguistics at the University of Halle, is an attempt to show that Homer’s language is firmly anchored in the Ionic vernacular of his time. While most of the book contains technical linguistic discussion which will deter many classicists, the theme should be of interest to Homerists generally. One of the most hotly debated issues in recent Homeric studies is the relationship between Homer and the Bronze Age. In this discussion, linguistics plays as large a role as archaeology, and discussions of the topic standardly include a reference to the vexed question of Myceneanisms in Homer. H.’s book addresses this question too and answers in the negative, which is hardly surprising in view of the author’s intention to emphasize the importance of recent Ionic elements.
The book falls naturally in two parts, chapters I to IV discussing archaisms and innovations in poetic languages generally and models of describing and explaining the Homeric language in particular. H. cites cases like Luther’s German, early 18th century Russian and the evolution of a Tocharian literary language in Central Asia in the 5th to the 8th century. The second part, chapters V to VIII, contains the linguistic discussions, focusing on the system of athematic inflection in the verb, which in Chantraine’s standard account1 gives the impression of chaos. H. argues that in many cases, ‘aberrant’ forms are not purely artificial creations, but find their place within a system of morphological innovation in eighth-century Ionic. The linguistic analyses presented are of uneven quality, but in general the author manages to stress the importance of the Ionic vernacular for Homer’s language.
As recognized in the preface, the field of Homeric linguistics demands a symbiosis of classical philology and Indo-European linguistics. But the author’s command of Homeric philology is sometimes deficient. It is surprising in these days to see stated on p. 1 (note 2): “Die Forschung ist sich einig, dass sowohl Ilias als auch die Odyssee Produkte des 8. Jh. sind (…)” While this is probably still the most widely held opinion, it can no longer count as a consensus; in the last few years dates from the 9th to the 6th century have been proposed,2 and the 7th century especially is gaining adherents. In the same way, on page 49, it is simply stated that Homer’s conception of the Iliad came only a few decades after the introduction of the alphabet in Greece and the founding of Greek ‘Texualität und Literarität’. Whatever one’s personal opinion on these questions, one would like to see them further discussed, not because every book on Homer must discuss every aspect of Homer, but because they are important for this book’s main theses. Today, of course, no one can be expected to command the entire field of Homeric studies, but for a book which stresses the importance of Homer’s own vernacular Ionic, the dating obviously becomes very important. And, in a book which claims that Homer’s many linguistic variants do not stem from an older epic tradition but rather from the effects of an emerging literary language, the relationship between the medial revolution and the Homeric poems would demand more treatment.
The first chapter lays out the preliminaries for the book. H. discusses the purported Mycenisms in Homer and concludes (rightly, to my mind) that they can all be explained in other ways. He discusses with sympathy the Berg-Tichy hypothesis3 that lines like ἀνδροτῆτα καὶ ἥβην (Il. 16.857 = 22.363) might reflect metrical archaisms, but then he rather bizarrely proposes that the short scansion of the first syllable in ἀνδροτῆτα might reflect a kind of muta cum liquida phenomenon. This is of course impossible, even when tacitly deleting the δ (H. puts it in parentheses) — which would in any case bring us far back in history, before the development of epenthetic d. The first chapter introduces the terms ‘anticipatory’ and ‘isolative’ anachronism. By anticipatory anachronism, the author intends forms in Homer that are linguistically progressive, and only later become normal. ‘Isolative’ anachronisms are innovations in Homer (and ex hypothesi in the Ionic vernacular of his time) which never became normal — historical ‘dead alleys’ so to speak. As they are per definitionem not much used outside Homer, it can often be hard to keep them apart from purely artificial forms. The recurring emphasis on anticipatory anachronisms is also problematic, since Hackstein most often compares ‘young forms’ in Homer with their Attic equivalents. It comes as no surprise that Homer’s forms can often be linguistically ‘younger’ than corresponding Attic forms, for it is well-known that Attic is more conservative than Ionic; witness for instance its preservation of the dual. It seems unnecessary to make so much fuss about ‘anticipatory anachronisms’ in treating this well-known phenomenon. Verbosity is a general problem with this book: the five pages on προθέουσιν, for instance, bring nothing new compared to what the old Ameis-Hentze said on the form,4 except for banalities like “Sprachlich individuelle Abweichungen (i.e. in this case the absence of the reduplication) sind als sprachliche Fehlleistungen bei emotionaler Erregung jedem bekannt” (p. 115). The second chapter brings a discussion of four different models for explaining the special character of Homer’s language, vz. Witte’s, Parry’s, Shipp’s and Crespo’s.5 Hackstein rightly concludes that these models tend to be too rigid in explaining the Homeric language on the basis of one principle, or many hierarchically ordered principles. On page 44, the author rashly brushes away explanations based on a preceding Aeolic phase:6 “So kann der Ablauttyp μέμηλεν, der oft als kunstsprachlich verkannt worden ist, als “sprachecht” und einem durchaus geläufigen, innergriechisch geneuerten Ablauttypus zugehörig identifiziert werden, […] Ein Erkenntnisgewinn für die Geschichte der epischen Sprachform ist möglich: Das flektierte Wort der frühepischen Dichtersprache verfügt über einen morphologisch homogenen ionischen Kern.” Now the recognition of unnoticed ablaut patterns is the main strength of Hackstein’s book, as we shall see, but nowhere does he prove that forms like μέμηλεν are specifically Ionic. It is not surprising that linguistic variation in Homer concerns the morphemes rather than the word kernel, for morphology is probably where you can find most prosodically relevant differences between the Greek dialects. In the third chapter, Hackstein brings examples from Luther’s German, early 18th Century Russian and the Tocharian texts from the Silk Road between the 5th and the 8th century, which show that morphological variation is a common feature of emerging literary languages. He stresses that such variants need not belong to different layers of a text. This is true, but Parry taught us that epic diction can have many layers even though the poems do not. We no longer have to posit an Aeolic original of the Iliad in order to account for Aeolicisms in Homer. But they must come from somewhere, and an earlier epic tradition is the obvious answer.
The fourth and last of the introductory chapters discusses three errors which in H.’s opinion have dominated Homeric studies, namely 1) use of the primary paradosis as the only authority 2) overestimation of the Kunstsprache and 3) use of historical morphology as a criterion for authenticity. While one might agree with H. on this, this book at times goes too far in the other direction, as we shall see.
On the methodological basis of these introductory chapters, Hackstein sets out to treat the athematic inflection in the different tense forms of the verb, present (ch. V), aorist (ch. VI), perfect (ch. VII) and pluperfect (ch. VIII). These chapters consist of case studies on specific forms and I will comment on some of them below, but first some general comments. The chapters are uneven, but much the best is the one on the perfect. H. shows that apparently unmotivated vowel alternations like ἀρήρει, ἄραρον belongs to a pattern of secondary ablaut that was productive at some point in the history of Greek. This demonstration is elegant and convincing, but here the problem of ‘isolative anachronisms’ surfaces. H. claims that this ablaut pattern is a specifically Ionic innovation. But as long as it does not appear in the later dialects, we simply cannot know from which dialect this innovation comes. We cannot know the quantities of Mycenaean a-ra-ru-ja and a-ra-ru-wo-a, and Hackstein is entirely right in dismissing the Cypriot evidence for an Achaean origin for the short-vowel forms, but as long as the Homeric poems contain Aeolic elements — and no one since Strunk 19577 has denied that — we cannot simply conclude that the forms in question are Ionic.
In the chapter on the perfect, H. reveals other innovations in the Greek apophonic series, which allow him to explain many obvious, but phonetically irregular etymological connections. Thus, renasalisation of the zero grade a in the ablaut series en – on – a, giving a new zero grade an can account for δήνεα < *dansea with a secondary zero grade to the root *dens. Another mechanism is revealed in the perfect. As is well-known, Greek sometimes substitutes long υ for etymological ευ as the full grade corresponding to zero grade short υ. Hackstein shows that the same analogy is at work in roots with a liquid or a nasal: secondary full grades are created by lengthening of the zero grade. Thus, beside the zero grade βαθ in βαθύς, we find the etymological full grade βενθ in βένθος and a new full grade (Ionic-Attic) βηθ in βῆσσα.
The chapter on the present is weaker. H.’s theoretical basis is the work of Bittner,8 which shows that verbs in German can be more or less strong, there being clear implicative structures, i.e. if the imperative is strong (shows umlaut), the rest of the paradigm does so too. If it is weak (no umlaut) the verb can still have a strong paradigm in the other forms, like werde! but present wirst. Furthermore, a strong 2./3. sg. of the present implies strong forms in the rest of the paradigm except the imperative, whereas a verb with weak forms here can still have a strong paradigm in the preterite, like hebt, hob. In other words, weak inflection first shows up in the imperative, then in the present and so on. H. tries to show that a similar process is at work in the abolition of the athematic inflection in the Greek verb, the imperative being the first conquest of thematic inflection, then the 2. and the 3. sg. active, followed by the third plural, the preterite singular and the medium. But the application of these schemes to Homeric verbs is not without its problems. Most verbs have both athematic and thematic forms beside each other. So beside the thematic 3. sg. ἱεῖ, τιθεῖ, διδοῖ we do find athematic ἵησι, τίθησι, δίδωσι. But in the second singular, there are no athematic forms beside thematic ἱεῖς, διδοῖς. In this case, the athematic forms would be metrically equivalent. So within a classical phasal model of the evolution of the epic language, we might suppose that at some time the tradition changed from a dialect that preserved athematic inflection in these verbs (Aeolic is a natural choice, as the athematic inflection was productive there) to a dialect without athematic inflection, like Ionic. In any case, it seems certain that metrical needs intervened with natural language evolution. This is not discussed by H., who seems to underestimate the force of the Kunstsprache in his eagerness to show that traditional Homeric philology has overestimated it.
The chapter on the pluperfect concentrates, not unexpectedly, on the verb ‘to know’ and especially the forms ᾔδει and ᾔδη. H. considers the latter form a hypercorrect version of the former, on the basis of the ongoing process of thematisation in the athematic verbs. The same procedure is supposed to have taken place in ἔσβη, allegedly hypercorrect for * ἔσβει from an old sigmatic aorist *esbese. The hypercorrect form could win through because of the analogical pressure from the other singular forms ἔσβην, analogical for ἔσβη from *esbesa and ἔσβης from *esebesas. This does not fit, however, with H.’s apparent claim (p. 117) that Attic knew thematisation of originally athematic imperfects only in the 3rd singular, i.e. ετίθην, ετίθης, ετίθει. Such a paradigm would, on the contrary, support an old * ἔσβει. In any case, the trend in Greek is clearly towards towards thematisation, and it seems improbable that hypercorrect forms won through in this verb alone in Attic. So, there seem to be no parallel cases for the supposed hypercorrection of ᾔδει to ᾔδη. Nor does H.’s derivation of this paradigm from an original root aorist convince. H. claims that the ablaut pattern (full grade in all forms except the 3rd plural ἴσαν) follows a pattern which is productive in the root aorists. As he rightly points out, following Hardarson,9 the root aorist generalized a non-alternating, often long vowel, which in certain root structures was shortened by Osthoff’s law in the third plural, i.e. ἔβην, ἔβημεν, ἔβαν. But there is nothing to indicate that such patterns were generalized into roots where Osthoff’s law did not work and thus nothing to recommend the idea that ᾔδει and ἴσαν reflect an old root aorist. This book largely consists of case studies on single words or paradigms and invites some comments on details (to be skipped by impatient readers!).
P. 99: the author gives the form κνέε without any indication that the form is reconstructed for the MSS’s κνῆ (Il. 11.639). The case for restoring κνέε is only presented twenty pages later, but it is not convincing. The author seems unaware that later Ionic has κνα.
P. 103: H. rejects the traditional explanation of 2. sg. εἰς‘you are’ as the original form εἶ + secondary ending -s. H. argues that the basis form εἶ is not attested in Homer or early Ionic lyrics. But it is (as H. acknowledges) the Attic form, and, since it is an archaism, Attic must have inherited it from the common Attic-Ionic dialect, which means that it must have existed in the prehistory of Ionic.
Pp. 105f.: H. follows Kiparsky’s derivation of the 3. sg. ending ‐εις from ‐ετι. He supports this sound law with Cretan πρειγυς, allegedly from *preti-g w h 2 -us, although both Kiparsky and Cowgill restricted this sound law to final syllables, a fact which would require at least a comment.
P. 132: The attribution of Sappho’s imperative ἔσσο (in the hymn to Aphrodite) to epic influence on the ground of the poem’s theme (‘thematisch zur Aufnahme von Epizismen prädestiniert’) is rash. Unlike, say, Sappho 44 (Voigt), this poem seems to contain no epic morphology, even line 11, which evokes Od. 2.151, contains specifically Aeolic forms. On the whole, H. seems to be unduly skeptical about Aeolicisms in Homer: commenting upon the kappa aorists (pp. 136ff.), he notes (p. 139) that ‐θετο (Il. 14.223) and θήκατο (Il 14.187) seem to be used indiscriminately within 40 lines. H. considers two possibilities, that the forms belong to different registers, or that the vacillation reflects a change in progress, but not the possibility that they belong to different dialects. The aorist without kappa could be Aeolic, as its combination with Aeolic apocope in the preverb κατ suggests. (The misprint ἐγκάθετο for ἐγκάτθετο obscures this.) If so, it must be archaic Aeolic, for in historical times the kappa forms have intruded at least into the 3rd plural.
P. 156 (and note 12): the author refers to his idea that a laryngeal disappeared in Indo-European when it was the second of four consonants, i.e. *CHCC > *CCC. In this case, one would like to see a discussion of -dhro- and -tro- derivations from set-roots, which always seem to show a reflex of the laryngeal. In some cases, this can be analogical from the corresponding verb (in much the same way that Latin aratrum has its long a from the verb arare), but in other cases such an explanation seems impossible, e.g. for Armenian arawr‘plough’: this suffix is not productive in Armenian, which in any case does not have a corresponding verb *aram, so it should rather be taken as the direct phonetic continuant of IE *h2erh3-trom with the laryngeal intact.
Pp. 164ff.: H. defends the paradosis’ συνοχωκότε (Il. 2.218) against the often-supposed συνοκωχότε, found in Hesychius and put in the text by West. H. follows Wackernagel and explains συνοχωκότε as the perfect participle to a denominative συνοχόω, with metrically conditioned active inflection. This verb does not exist, but H. claims that perfect participles can have adjectival function beside nominal stems, without any intermediating verb. H. illustrates this claim with German ‘gewillt’ directly to ‘Wille’, but does not give any Greek examples. Indeed, there are cases like βεβροτωμένα (Od. 11.41) and τεθυωμένον (Il. 14.172). But these are formed from relatively common epic words, whereas the supposed nominal basis σύνοχος is not found in Homer. Indeed, σύνοχος is found in the required sense only in (and not from, as H. writes) Euripides. Later attestations come from medical authors and have the different meaning ‘unintermittent’.
Pp. 246ff.: H. discusses the anomalous ending – -̔σ̓θε of the 2. pl. active in the perfects ἐγρήγορθε, ἄνωχθε, πέποσθε. He believes that the forms come from * -̔σ̓τε (analogical after the 2. sg. ending ‐σθα), which became -(σ in roots ending on liquid, nasal or ‐θ, and this variant was subsequently generalized. But it seems unreasonable that Greek should have gone through so many analogical steps to create an active ending which would be identical to the medio-passive ending. Note also that the evolution *-rst- > *-rht- > –ρθ, though not unreasonable, cannot be supported by Klingenschmitt’s interpretation of παρθένος as *pr-stenos (with sonant r): this etymology must be rejected, because the word is well-attested in dialects were we would expect ορ from sonant r.
Although I disagree on many details in H.’s linguistic analyses, I would like once again to stress that there are quite a few convincing explanations to be found in this book. In particular, the author manages to demonstrate the existence of systematic patterns in innovations that up to now have appeared unsystematic. But sometimes H. seems to be carried away by his own desire for classification. He goes too far when he (pp. 172-178) divides the perfect classes in Homer into three main classes, each with three or four subclasses, some of whom have further subclasses. And claiming (p. 175) that the form δείδια (with zero grade in the root as opposed the o-grade in δείδω) shows a beginning change from class I.2 to class II.1 is a parody when one considers that both class I.2 and II.1 have only one member.
The many indices (subjects, cited words and passages) make the book easier to use, but the bibliography could have been better handled. Some items which are cited in the text are missing in the bibliography (especially Peters 1994 from FS Otakar Klíma would be hard to find for the non-initiated). Works by the same author which appeared in the same year are not differentiated. (Even the second volume of Valentin Kiparsky’s Russian historical grammar and Paul Kiparsky’s article ‘A Phonological Rule of Greek’ are both cited as Kiparsky 1967.) Because of its flaws, the book must be read with a critical mind — but, still, it should be read by all scholars of Homeric linguistics. Other Homerists will probably find it tough going. The main thesis of the book, its stress on the Ionic vernacular, might seem less sensational than the author thinks, but there are many interesting and convincing explanations to be found here.
[[For a response to this review by Olav Hackstein, please see BMCR 2004.03.12.]]
Notes
1. Grammaire homérique (Paris 1958), first volume, chapter 22.
2. 9th Century: Ruijgh C., “Sur la date de la creation de l’alphabet grec”, Mnemosyne, 51 (1998), p. 682. 6th Century: Jensen, M. S. Symbolae Osloenses 74 (1999), p. 34. A 7th century date has recently been favoured by influential scholars like Burkert, Kullmann and West.
3. See Berg, N. and Haug, D. in Symbolae Osloenses 75 (2000), pp. 5-23 for discussion and references.
4. Ameis-Hentze-Cauer, Homers Ilias, für den Schulgebrauch erklärt, Leipzig 1894-1913 on Iliad 1.291.
5. Witte, K. (1913) in Realenzyklopädie der klassischen Altertumswissenschaft, vol. 8: 2213-2247. Parry, M. (1932) in Harvard Studies in Classical Philology, 43:1-50 (= The Making of Homeric Verse, pp. 325-364), Shipp, G. P. (1972, first edition 1953) Studies in the language of Homer, Cambridge, and Crespo (1997) in Létoublon, F. (ed.) Hommage à Milman Parry, Amsterdam.
6. Wathelet’s 1970 study Les trait éoliens dans la langue de l’épopée grecque is completely missing from this book.
7. Strunk, K. Die sogenannten Äolismen der homerischen Sprache, Diss. Köln.
8. Bittner, A. (1996), Starke ‘schwache’ Verben – schwache ‘starke’ Verben: deutsche Verbflexion und Natürlichkeit, Tübingen.
9. Hardarson, J. A. Studien zum urindogermanischen Wurzelaorist, Innsbruck 1993, pp. 151f.