The original objective of the Ph.D. project that eventually grew into this book was rather grotesque: I planned to establish a relative chronology of all sound changes that took place between Proto-Indo-European, Proto-Greek, and the different dialects of Ancient Greek. After some time, it became clear to me that the syllabic liquids constituted one of the most important problems. There is a large variety of different and contradictory opinions on their development and reflexes in the Greek dialects. Moreover, it was a topic with potentially large consequences, not only with respect to questions of relative chronology, but also for the genesis of the four main dialect groups of alphabetic Greek. When a new possibility to tackle the problematic double reflex of *r̥ suggested itself to me, I started to put all my time and effort into this problem, and eventually decided to devote the entire dissertation to this topic.
When I defended the dissertation in December 2013, I was convinced (as I still am) that the main results were sufficiently plausible and innovative to be published.1 However, as I was not completely satisfied with the presentation of my arguments, and as I felt that my work would improve if it got the chance to ripen, I continued working on it off and on. The present book emerged from the dissertation by a gradual process of revising, expanding, deleting, and weighing the arguments over and over again. Below I will comment on the differences with the 2013 dissertation in more detail. In terms of conclusions and the main arguments, however, the works are very similar.
An important thread running through this book is my claim that *r̥ and *l̥ did not have a double reflex in any Greek dialect. Of course, this claim will not surprise anyone trained in Neogrammarian principles, nor am I the first to make it. Nevertheless, the contrary view that the place of the anaptyctic vowel beside *r̥ and *l̥ varied per word, or at least that the distribution between both reflexes still remains to be found, has become widespread in handbooks concerned with the historical phonology of Greek.
Apart from a strict adherence to the principle of regularity, another important methodological point applied throughout this book is a search for the concrete models and motivations that may (or may not) underlie the analogical reshaping of word-forms or the derivation of new lexemes. In my view, we must not be satisfied with the observation that a certain analogy is possible on paper; it is often possible to go further than this and to tell whether an analogy is plausible or implausible, paying attention to the contexts in which a word is used. This holds especially for the language of epic, where it is often possible to indicate a concrete impetus for the reshaping or creation of a specific word, phrase or formula.
The focus of this work is on reflexes of syllabic *r̥ and *l̥ that can be reconstructed for Proto-Greek. Developments involving PIE laryngeals (such as the problematic double outcome of *CRHC clusters in Greek) are touched upon, but they do not occupy center stage. My basis is a thorough and up-to-date etymological discussion of all words containing a reflex of Proto-Greek *r̥ or *l̥, in all dialects of Ancient Greek, including Mycenaean. A fair number of new etymologies and novel reconstructions of forms are presented (for an overview, see section 12.1). In addition, there are various fresh discussions of issues in derivational morphology, especially concerning the ‘Caland system’.
The main innovative hypothesis advanced in this work is the bold claim that ‑αρ‑, rather than ‑ρα‑, is the regular reflex of word-internal *r̥ in Ionic-Attic. This idea first took a rudimentary shape when I realized that two seemingly independent metrical peculiarities of Homeric Greek had to be related: on the one hand, Wathelet’s (1966) observations about the distributions and origin of muta cum liquida in Homer; on the other, the realization that the onset of κραδίη rarely makes position in Homer, an oddity that had been observed earlier by Hoenigswald (1991). In combination, these two peculiarities suggested to me that *r̥ had been preserved until not too long before Homer.
The main breakthrough took place when I realized that ‑ρα‑ in various exclusively Homeric words could be seen as an artificial reflex of word-internal *r̥, and as such could be contrasted with the regular reflex ‑αρ‑ in words that occurred in Ionic and Attic prose. It appeared that a prolonged retention of *r̥ in the Dark Age epic tradition could explain both the reflexes ‑ρα‑ and ‑ρο‑ in words that are virtually limited to Epic Greek and the peculiar prosodic behavior of many such words.
Along with this novel hypothesis came a large number of problems. All Ionic and Attic forms with ‑ρα‑ had to be accounted for, especially those forms which do not appear in Epic Greek. Moreover, the evidence from all other Greek dialects had to be re-examined. In this domain, too, the attempt to determine the regular slot of the anaptyctic vowel proved fruitful. For instance, it appeared that ‑ρο‑ was the regular, unconditioned reflex in Aeolic dialects; that a regular reflex ‑ro‑ could be excluded for Mycenaean, and that Cretan normally has ‑αρ‑, but a conditioned reflex ‑ορ‑ after labials.
My ideas concerning the reflexes of *r̥ that are specific to epic have led me to propose a new model concerning the relationship between Epic Greek and the poets’ vernacular(s). The consequences of this new model for the prehistory of the epic tradition are still difficult to oversee. From 2016 to 2019, I have been elaborating and testing this model within the project ‘Unraveling Homer’s Language’. An important result corroborating the model proposed here is my analysis of the formulaic and metrical behavior of the Homeric verb ῥέζω ‘to do’. I have decided not to incorporate this result in the present book, but to publish it separately in the volume Language Change in Epic Greek and other Oral Traditions, which is to appear in the LSIE series.
As said above, the present book differs in many details from the 2013 dissertation, and the reader who compares both works will see differences in formulation in almost every paragraph. In this sense, I have completely revised the book. On the other hand, these revisions hardly ever affected the core of the argument. It is my hope that they have made the whole more persuasive and more clearly presented. I have added many bibliographical details (including secondary literature published after 2013), expanded and refined various etymological treatments, and elaborated certain new ideas with more precision (such as those on the comparison of Myc. to-ro-no-wo-ko with Hom. θρόνα, now in section 2.5.2, originally in 2.2.1). I left out a couple of digressions that were not relevant to the main argument and which required more extensive argumentation, for instance:
-
the accentuation of feminine stems in *‑i̯a (Van Beek 2013, section 4.1.1);
-
a new proposal concerning the etymology of καρτερός (Van Beek 2013, section 5.3);
-
a new proposal concerning the etymology of εὐνή (Van Beek 2013, section 11.3.1), which I have now elaborated in Van Beek fthc.
I have changed my opinion on one important point: the presence of Aeolisms in Epic Greek. While finishing my dissertation, I was impressed by the insight that Homeric forms with ‑ρο‑ did not have to be Aeolisms, but instead could also show a vocalization of artificially retained Epic *r̥. While I still believe this to be true in many cases, it does not follow that other forms with *r̥ > ‑ρο‑ (or ‑ρα‑: see section 8.4.3) cannot be Aeolisms. Moreover, it was rash and unnecessary to exclude the presence of Aeolisms in Homer generally.
New features of this book compared to the dissertation include:
-
a discussion of the Greek evidence for an u-reflex of syllabic liquids (section 1.3.2);
-
an introductory section on Epic Greek as an artificial linguistic form, including a brief discussion of theories on the origin of the hexameter (section 1.5);
-
the addition of more Mycenaean evidence and of more secondary literature on this dialect (chapter 2);
-
the recognition that certain a-spellings in Mycenaean must be taken seriously (section 2.4);
-
a brief treatment of the evidence from Argolic (section 3.2.4);
-
the realization that the analogical root vocalism of κρατύς ‘strong’ and other adjectives in ‑ύς can be ascribed to inter-paradigmatic rather than intra-paradigmatic levelling, for instance after the forms of comparison (section 4.3.3);
-
a quantitative analysis of the evidence for muta cum liquida in Homer (section 6.5);
-
the addition of ἀταρπός / ἀτραπός ‘path’, ἄρχω ‘to be first’ and χάρμη ‘fighting spirit’ as serious pieces of evidence for the regular vocalization of *r̥ (section 9.6);
-
clearer etymological discussions of many words with λα, such as πλάξ, λάγνος and λαγωός (chapter 10).
The dissertation became available online by June 2017.