### Ahuvia Kahane, The Interpretation of Order. A Study in the Poetics of Homeric Repetition. Oxford Classical Monographs. Oxford: Clarendon Press, 1994. Pp. 190. \$39.95. ISBN 0-19-814077-0.

Reviewed by M. Steinrueck, Lausanne.

If verse is a transformation of phrastic structure, then K. has written a sort of semantics of the Homeric hexameter. By means of statistics (with lists at the end of the book) and literary interpretation he establishes relationships between morphology, metrical position and the semantics of the narrative context.

The first chapter establishes a relationship between "metrical units" and "sense units". In fact, there are positions in the meter where words (or word groups) finish much more regularly than elsewhere. These positions, described already by Fraenkel, Porter and O'Neill, create a division of the hexameter into two cola divided each into two sub-cola. It is this fourfold structure which makes it possible, on three levels, to establish a link between metre and semantics.

a) It shows that the formula-definition conceived by Parry is too rigid. The small metric units also enable one to deal with words only. Visser's theory of versification by nucleus and peripheric elements can support K.'s point of view.

b) The fourfold metrical order also justifies the link between word-order and metre: initial position = first colon, internal position = second and third cola (this lack of distinction is nowhere explained), terminal position = fourth colon.

c) From frequency and lack of word-boundary on certain positions K. derives a rule of "static" norm and "dynamic" deviance. This rule enables him to do wonderful readings pointing out the "showing" of many passages (for instance Od. 5.1f.): when static contents coincide with the (normal) static distribution of word-boundaries or dynamic contents are superposed by the dynamic (unusual) position of word-boundaries, we can say that the rhythmical shape "shows" what we are told by the content.

The second chapter begins by evoking the pathos of the first words of the Iliad and the Odyssey; every time the three conditions determining MH=NIN and A)/NDRA are fulfilled -- i.e. initial position of the word, accusative and reference to the main epic theme (Achilleus' wrath, Odysseus) -- K. speaks of an accusative-theme-word-pattern. Indeed, the word MH=NIS, and, what is more surprising, the common word A)NH/R usually mean something else when they are not accusatives or when they occur in the middle or at verse-end. The same point is made with NO/STOS. Two instruments are used:

a) In order to indicate clearly what points to the "theme" and what does not, K. distinguishes from the simple lexical "denotation" (e.g. A)NH/R = man, husband), the "anaphoric" reference to the narrower context (A)NH/R = the man the text just mentioned, cf. Od. 3.24: Telemachus) and the "deixis", which, within a broader context, determines one special thing or theme (e.g. A)NH/R = "the man of this story").

b) In order to emphasize the contrasts and to localize the exceptions, K. uses four categories of items: first all accusative theme word patterns (initial position, accusative, directly pointing for example to Achilleus), second all accusatives of the word in initial position too but pointing indirectly to Achilleus; in the third category the nominatives and genitives are at the same time positioned in the middle or at then end of the verse and they refer to a person other than Achilleus; the fourth category indicates the few (!) exceptions: a non-initial position pointing to Achilleus.

Semantics of position do not rely on the universal acceptance of the significance of "beginnings" but rather on the mere opposition of marked and unmarked items. Therefore, in the third chapter dealing with the proper name vocative, it is not necessarily initial position which has to be marked. The initial vocatives, however, represent the majority of addresses to minor characters and their position is regarded as a default mode, a neutral pattern, while the terminal position coincides with addresses to protagonists. So, terminal position is marked. The internal one -- it has often been said -- is the place proper of apostrophe, a figure (cf. appendix IV) which indicates the narrator's (or the audience's) sympathy. This sympathetic pattern is used for Menelaos but not for the protagonist Agamemnon (usually addressed in terminal position); Patroclus and Eumaeus have this type address while the protagonists Achilleus and Odysseus have none.

a) The main argument is perhaps the opposition of a vocative's metrical shape and its position. MENE/LAE, for instance, could very well be placed at the end of the hexameter, but in an apostrophe we find him in the middle of this verse. On the contrary, the following chapter tends to exclude metrical impact.

The fourth chapter widens the frame of the third and attempts to show that even a rather "mainstream pattern" such as the proper name nominative can be a mark when positioned at the end of the verse. It is, then, a typical heroic feature just as Colt and Stetson mark a Western hero. This corpus is not easy to classify; perhaps the most important point of reference is the score built by the two main heroes and their deities. Not only do Odysseus, Achilleus, Athena and Apollo build one thematic net, they also have in common a double shape (one with one consonant and a second with two consonants) able to be placed at the end and elsewhere.

In general, I have enjoyed K.'s reading of passages (e.g. p.105 the terminal position of MENE/LAE linking this hero in the fourth Odyssean book with Odysseus) within the corpus of these four chapters; however, there are certain questions: just to name one, why does Telemachus in Od. 3.124f., p.63, point to Odysseus, while he does not do so in Od. 3.160, p.100? There is a narratological argumentation (unfortunately a little dispersed in the book) I find interesting to discuss. One could put together the arguments as follows: positioning in the hexameter is a mark (based on the feature "marked-unmarked") that presents a subject's particular aspect (as the main theme, as the protagonist, as the hero). So, it is one type of the "deiktika" revealing the perspective of the speaker. Now, in an epic plot, there are characters and a (?) narrator having each a different point of view. But the positioning-marks are the same in (Odyssean or Iliadic) narrative parts as in speeches. A contradiction? K.'s answer is "no" and his explanation the following: both the narrator's discourse and the character-speech are shaped into hexameters, but only the narrator's discourse could even be defined by its hexametric diction. On the contrary, e.g. the character Odysseus is never supposed (by the audience) to address the suitors by hexameters. K. draws the conclusion that a positioning mark even when surreptitiously introduced into the character's speech always belongs to the narrator's perspective. A narrator, seemingly, does not share the poet's biography. However, according to K., the words they pronounce and say respectively are the same, and both instances of speaking can be subsumed by one focalization/perspective (called NFI according to De Jong 1987).

Now, it seems to me that this model creates certain problems in some of the passage discussions dealing with the constraints of the system limiting the free choice of a perspective (a mark). For instance, the frequent terminal position of some vocatives could be simply explained by the same position of these names in their nominative shape: "No matter how simple the causes of certain phenomena, they can still have functions and effects ..." (p.86). So, where the narrator has a choice (between positions), K. claims for the narrator's point of view, but where he has not, the effect on the audience of a even unintentional position mark cannot be avoided. Who, then, is speaking? The answer might be much easier given by the ancient narrator-concept consisting of the Muse (the system?) speaking through the mouth of an unknowing A-oidos.

This book is easy to read, interesting and innovative.