Re: Metrical Stress, Feet, etc.

On Friday, February 6, 2004, at 05:52 AM, David Peterson wrote:

> To Ray Brown:
>
> If stress had nothing to do with pre-classical Latin, how does one
> explain that word-final bi-, and even tri-moraic syllables don't get
> stressed?
I don't understand the question, since:
(a) I certainly did _not_ say that stress has nothing to do with
pre-Classical Latin nor do I believe this to be so;
(b) I do not see by what logic the conclusion follows from the condition
in any case.

There seems to be some misunderstanding so, hopefully, the following will
be helpful.

Part 1: The condition (If stress had nothing to do with pre-classical
Latin)
--------------------------------------------------------------------------------
I most certainly, as I've said, do not claim this. In my mail of yesterday
I wrote:
"Stress, whether primary or secondary, is irrelevant to moraic based
metrics of, e.g. Classical Greek and Latin."
I was talking about _Classical_ Latin, not pre-Classical or post-Classical
in which (generally) stress most certainly did play a part. Actually, I
did overstate the case a little, as I shall show below. As far as ancient
Greek metrics are concerned, I stick by what I said; but in Classical
Latin metrics things were not so clear cut and, in fact, stress did have
an effect - but _not_ in the way in which it operates in languages with
stress-based rhythms.

The only examples of pre-classical verse are AFAIK the remaining fragments
of so-called Saturnian verse. How this is to be interpreted metrically is
debatable but I was under the impression it was generally thought to be
stress-based, perhaps rather like G.M. Hopkins's 'sprung rhythm'.

Although some scholars still, I believe, claim classical Latin had pitch
accent like ancient Greek certainly had, the vast majority of scholars are
certain it had stress accent. Further, the general consensus is that in
early Latin word accent was on the first syllable (as, e.g. in Hungarian,
generally in the native words of germanic languages, in Gaelic etc.) and
that at some time before the Classical period it changed to the well-known
Classical rules:
disyllabic always on the first syllable, other polysyllabics:
- on the second-last syllable if that syllable is heavy,
- otherwise on the third-last syllable whatever the quantity of that
syllable.
(Quenya follows the same rues :)

The Classical Latin rules were certainly those by and large of 'Vulgar
Latin' (i.e. proto-Romance) as we can tell from developments in the
Romance languages. There is also evidence from Latin prosody itself, as I
shall show below.

As for the initial syllable stress of pre-Classical Latin, one has only to
consider the weakening of vowels after the initial syllable in compounds,
e.g.
ad + facio --> afficio
ad + factum --> affectum
con + claudo --> concludo
ob + caedo --> occido
etc.

That change of habits by speakers of stressed-based languages do occur is
attested elsewhere, e.g. Middle Welsh had final syllable stress whereas
modern Welsh has stress on the second-last syllable.

But the point is that Latin has always, as far as we can tell, been a
language with word-stress and, I suspect, popular verse was always
stress-based.  The pre-Classical Saturnine verse was probably stress-timed
(tho the details are unclear) and the poetry of late & medieval Latin was
stressed-based.

Classical Latin verse, however, is a conscious imitation of Greek norms.
Now ancient Greek was a very different matter. There is at no period any
evidence of vocalic weakening in 'slack' syllables or any other indication
of word stress. All the evidence points clearly to pitch accent. The
ancient Greek meters were 'quantitative' or moraic. There were two types
of syllable:
i. light (monomoraic), the syllabic nucleus was a short vowel and the coda
is zero.
ii. heavy (bimoraic) - all others, i.e. syllables with codas, and
syllables with zero coda but long vowel nucleus.
(It might be thought that long-vowel nucleus + non-zero coda would
constitute an 'extra heavy' or 'trimoraic' syllable - but this was not so)
.
We may think of light syllables as being equivalent in music to a quaver
(eighth note) and a heavy syllable to a crotchet (quarter note).

Now syllable-timed meters were admirably suited to ancient Greek. They
were not suited to Latin for two reasons:
i. All the evidence, as I've said, points to Latin being a language with
word stress;
ii. Latin has a much higher proportion of heavy syllables than Greek had
which, in fact, did cause problems in the adapting of some Greek metric
forms.
Nevertheless, the educated Romans did adopt and adapt the Greek meters.
Their verse must have sounded very strange to the uneducated.

But in adapting the meters, the Latin stress could not be entirely
disregarded. Of particular difficulty was fitting the lyric Aeolian meters
to Latin and many struggled. It fell to Horace to do the thing
successfully and he set the de_fact Roman norm for these meters. But the
Sapphics and Aeolian meters of Horace have a very different feel from the
lighter meters of Sappho and ancient Greek poets of Aiolia.

But I'll confine my remarks now to what, I guess, is the most well known
of the ancient meters: the dactylic hexameter. The basic 'foot' (bar) was
the 'dactyl':heavy-light-light, and the hexameter, as it's name implies,
consisted of six of them.
In the first four feet, the dactyl might freely be replaced by a spondee:
heavy-heavy. But the last two feet were fixed: the fifth being a dactyl
and the 6th a spondee.

Now it would be natural reading such verse to put a slight stress on the
heavy syllable that began each foot. It is notable that, with few
exceptions, the word and rhythm stress coincide in the first foot (often
unavoidable, even if not intended) and in the 5th and 6th feet (not
necessary, and clearly done deliberately). In the second, third and fourth
feet there word stress ran counter to the verse rhythm. Clearly, the
Romans felt such counterpointing could not continue to the end of the line
without losing the rhythm altogether. How, then, should one read Classical
Latin verse?

School kids normally chant such verse (if at all now), ignoring the word
stresses and just going along with the beat, so to speak. That's what we
did when I was at school and that, according to IIRC Maurizio Gavioli, do
  Italian schoolkids - and I suspect it's pretty common elsewhere. Now
Sidney Allen also suggests (in Vox Latin) that the Romans did this as well
and that this formed a 'counterpoint' with natural word stress of the
ordinary language.

However, it seems to me that such counterpoint is really just a 'ghost'
and the coincidence of word-stress and rhythmic beat in the final two feet
is rather pointless as the rhythm is, in fact, never lost.

The alternative is to read the hexameters with their ordinary word stress
which, of course, run counter to the verse beat in the 2nd, 3rd and 4th
feet. If one is careful to pronounce long vowels as long vowels and
geminate consonants as geminate (i.e. not indulge in anglophonic habits),
the result in my opinion is far more interesting than the tum-tity
tum-tity of the method above. One gets _real_ counterpoint with the
rhythms coming back together at the end of each line and the effects are
very interesting. I am quite certain in my own mind that this is what the
Romans actually did and that the poets exploited this feature in novels
ways that, of course, the Greeks never could.

I did once start experimenting with some of Catullus' verse to see the
effects of this counterpoint; the results were very interesting. In my
opinion this is an overlooked feature of Classical Roman prosody and if I
had both the time and resources it is one I'd like to research.

In other words, it is my belief (and I know this is controversial) that in
Classical Latin verse:
- stress played no direct part in the rhythmic beat of verse (which was
moraic, following the classical Greek patters);
- but that stress did play a part in producing counterpoint.

ObConlang: an interesting idea would be to develop a verse form for, say,
conlang A which was basically unsuited for it and was borrowed from
conlang B for which it was eminently suited. The speakers of A, however,
exploit B's model in ways the speakers of B never could  :)

Part 2: The conclusion (how does one explain that word-final bi-, and even
tri-moraic syllables don't get stressed)
------------------------------------------------------
I don't understand how this logically concludes from the condition. If
stressed played no part then no syllable would get stressed whether word
final or not - the position's irrelevant.

But, as I've said, it's the opinion of most, and certainly of me, that
latin was a word-stress language. Now it has been observed that languages
with word stress come in two varieties.  Some like Russian and modern
English can place their stress on practically any syllable, including the
final. In such languages the stress tends to be strong. Other languages
languages have the stress fixed; the common position in natlangs seem to
be word-initial, word-final (less common - I believe Hebrew does this),
the second-last (e.g. Polish, modern Welsh) and the Quenya-Latin system.

But interestingly the development of the definite article in the western
Romancelangs do quite clearly point to the stressing of 'ille' and 'ipse'
on their final syllables when used proclitically.
---------------------------------------------------------

But I think I've said enough & the metrics that interest me and, I guess,
David.  I hope this long screed has been helpful.

Ray
===============================================
http://home.freeuk.com/ray.brown
ray.brown@freeuk.com    (home)
raymond.brown@kingston-college.ac.uk   (work)
===============================================
"A mind which thinks at its own expense will always
interfere with language."         J.G. Hamann, 1760