Form and Content. Conscious or Unconscious?

Introduction
In this paper, I would like to discuss the relation
between content and rhythmical form in Euripides’ Orestes. My study of work-internal differences
in Euripides’ rhythmical style is compared to the
earlier study by Philippides (1981), and a large
part of the discussion will concern the similarities
and differences between her results and mine. An
explanation for these differences will be sought in
the different nature of the phenomena studied and
in the different methods used to obtain the results.
Furthermore, I shall demonstrate that the relation
between form and content raises a question that is
important for another branch of Euripidean stylometry, viz. the study of the chronology of his
works, as well as for chronology (and attribution)
studies at large.
Background
The rhythmical style of Euripides (485-406 BC)
has been the object of study for almost two centuries now. Particularly the development in his use
of the iambic trimeter has been studied extensively. The iambic trimeter, in its basic form a 12-syllable line, is the main verse type of the spoken
parts of Greek drama. It turned out that Euripides’
use of resolution (the representation of one verse
element by two short syllables) increases gradually from his early to his late tragedies and can,
therefore, be used for the (approximate) dating of
his undated tragedies. Recently, Euripides’ use of
resolution has also been studied work-internally.
Philippides (1981) studied six of Euripides’ tragedies in this respect, three early and three late ones,
and her conclusions are quite interesting: a significantly higher incidence of resolution “appears to
coincide with scenes of dramatic intensity”, and a
significantly lower incidence “often coincides
with non-excited passages, narrative, rationalizing
discourse and scenes of camouflaged plotting or
lying” (both p. 107). It seems that the repeated
occurrence of the two short syllables of resolution
reflects in some way the excited mood of a passage. The rhythmical form, then, is found to support
the content.
My own study of Euripides’ work-internal style
focuses on elision. Elision occurs when a short
final vowel is followed by a word that begins with
a vowel. The elision of a vowel entails the loss of
a syllable for the rhythm: the syllable containing
the elided vowel merges with the following wordinitial one. In a way, elision and resolution have
the same effect: they both offer a way to add an
extra syllable to the line. Therefore, it seems possible that a concentration of elision has the same
function as a concentration of resolution, viz. to
indicate a situation of excitement. The findings of
Olcott (1974) for Sophocles seem to support this
hypothesis. She notes that the type of elision that
occurs on a caesura coinciding with “a break in
sense” is found “in unusual circumstances, that is,
emotional or emphatic” (p. 61). It will, therefore,
be my aim to discover whether differences in
elision use within Euripides’ works can be linked
with differences in content; and if that seems to be
the case, to see whether the results are comparable
with those of Philippides. For this, I am mainly
looking at two of Euripides’ plays: the Medea (431
BC), one of Euripides’ early plays, and the Orestes
(408 BC), one of the late plays. Here, I shall
discuss only results of the analysis of the Orestes
(also part of Philippides’ corpus).
Method
The elisions (and resolutions) are counted with the
help of a database system, designed for my project
by Jan R. de Jong (Computer Centre of the Faculty
of Arts, University of Amsterdam). The input of
the database consists of the electronic text of Euripides (from the CD-ROM of the Thesaurus Linguae Graecae) and a rhythmical analysis, based on
that electronic text and provided by a parser developed for the purpose (see De Jong and Laan,
1996). The rhythmical analysis is necessary to find
the resolutions, and also for the study of the use of
certain types of elision (for instance, elisions occurring at certain verse positions, e.g. the caesura).
To get some measure of objectivity about the
significance of the elision (or resolution) counts,
one needs statistical methods. As my data are
sequential, i.e. occurring in a series of trimeters, it
seems logical to adopt a model for time series data.
On the advice of prof.dr. Jan G. de Gooijer (Dept.
of Economic Statistics, Univ. of Amsterdam), I
use Change-point analysis, a method that looks for
the point of change in a series of data. A programme developed by him looks, in every series (i.e. in
every uninterrupted sequence of iambic trimeters,
e.g. an epeisodion or ‘act’), for a change in elision
(or resolution) use, either a change in mean or a
change in variance. The Orestes has six trimeter
series that are long enough for such an analysis.
Results and Comparison
The results are promising. A significant changepoint (already reported in Laan 1995, p. 275) was,
for instance, found at Orestes line 1030, with a
high mean of elision before the change-point and
a lower mean afterwards. This change in elision
use clearly coincides with a change in mood, from
utter despair to a calmer frame of mind. A concentration of elision, then, seems to have a function
comparable to a concentration of resolution.
When my results are compared with those of Philippides, it turns out that, so far,
(1) in the case of a significant change that involves
a high amount of elision, the high part never
coincides with a passage significantly high in resolution, but
(2) in the case of a significant change that involves
a low amount of elision, the low part sometimes
coincides with a passage significantly low in resolution.
The non-coincidence of highs of elision and resolution may be explained in several ways:
(1) both elision and resolution add an extra syllable to the line. A concentration of both might give
too much information compressed in too few lines.
In this case, elision and resolution are just two
ways to get the same effect, which cannot be used
together.
(2) resolution and elision might differ in their
dramatic function. Elision might, for
instance, be used to indicate a different kind of
emotion than resolution.
(3) a combination of both.
The difference between Philippides’ statistical
method and mine may also have some effect on
the difference between our results. If, for instance,
a whole trimeter series is consistently high in
resolution (or elision), my method (see Method)
would not detect any significant change, since
there is none in that particular part of the play; in
such a case, however, Philippides’ method might
find that (almost) the whole passage is significant,
since it computes the significance in relation to the
amount of resolution in all trimeters of the play
(Philippides, 1981, pp. 49-55), and not just in
relation to the amount of resolution in that particular sequence, as my method does. I.e., my method works on the level of a trimeter series, hers
on the level of the whole play. Moreover, I am
looking for a point of change in resolution (or
elision) use, while Philippides is investigating
whether a fixed set of lines shows a high (or low)
amount of resolution. This raises questions about
the relationship between the results of the two
methods.
Consciousness
Despite the differences just described and although the findings for elision still need corroboration, it seems justified to conclude that Euripides’ use of elision, like his use of resolution, seems
to be linked in some way with the content. This
relation between content and form appears to indicate that Euripides made a conscious use of
resolution and elision. However, the spectacular
increase in resolution from his early to his late
plays has generally been deemed an unconscious
phenomenon (see e.g. Kitto 1939, p. 179), and the
less spectacular increase in elision (see Laan 1995,
pp. 274-275) would perhaps also be considered
unconscious by some. And unconsciousness has
actually frequently been taken as a necessary requirement for a phenomenon in order to be considered a valid criterion for dating (or attribution);
see for dating e.g. Ceadel 1941, p. 67, and for
attribution e.g. Holmes 1994, p. 88.
I think, however, that the increase of resolution is
still valid as a dating criterion, despite the indication that Euripides made a conscious use of resolution within his works. For, as I pointed out
before (Laan 1995, p. 273f.), Euripides’ stylistic
development is an exemplary case, because of the
availability of securely dated works – with a range
of dates from early to late – that clearly show a
gradual development; i.e., his case is exemplary
regardless of the unconsciousness of his resolution
use. We should, then, ask the general question
whether unconsciousness always is a necessary
feature for phenomena that are used as dating (or
attribution) criteria? May it not be possible that
there are cases, such as that of Euripides, where
conscious phenomena do not have to be excluded?
Conclusion
All in all, it seems clear that the study of work-internal differences in Euripides’ rhythmical style
provides us with a little more knowledge of his
dramatic technique. Moreover, Euripides’ apparently conscious use of resolution and elision within his works raises doubts about the unconsciousness of his resolution and elision use in general.
And as resolution still seems a valid criterion for
dating, this forces us to reconsider the requirement, often made in stylometric studies, of unconsciousness for phenomena used as dating or
attribution criteria.
References
Ceadel, E.B. (1941), “Resolved Feet in the Trimeters of Euripides and the Chronology of the
Plays”, Classical Quarterly 35: 66-89.
174
Holmes, D.I. (1994), “Authorship Attribution”,
Computers and the Humanities 28: 87-106.
Jong, J.R. de, and Laan, N.M. (1996, forthc.), “A
Grammar for Greek Verse”, in: Nancy Ide and
Susan Hockey (eds.), Research in Humanities
Computing 4. Clarendon Press, Oxford.
Kitto, H.D.F. (1939), “Sophocles, Statistics and
the ‘Trachiniae’ ”, American Journal of Philology 60: 178-193.
Laan, N.M. (1995), “Stylometry and Method. The
Case of Euripides”, Literary and Linguistic
Computing 10.4: 271-278.
Olcott, M.D. (1974). “Metrical Variations in the
Iambic Trimeter as a Function of Dramatic
Technique in Sophocles’ ‘Philoctetes’ and
‘Ajax’”. University Microfilms, Ann Arbor.
Philippides, D.M.L. (1981). “The Iambic Trimeter
of Euripides. Selected Plays”. Arno Press,
New York.