latin grammar: reflexive pronoun + impersonal

[Adam]

Hello all,

First just want to say that it's too bad there's so much crap being
posted to this newsgroup. Is there another, less cluttered newsgroup
where people are posting legitimate questions?

I have a question about the pronoun "se" and its use in a subordinate
subjunctive clause or indirect statement. According to my book (which
is really quite excellent!) unless modified by "ipse", occurances of
forms of "se" in a subordinate clause or indirect statement will refer
to the subject of the main clause. But how does that work if the main
clause is an impersonal such as "oportere"? One of the exercises in my
book is the following:

Pacis petendae causa ducem oportebat polliceri se nemini post bellum
nociturum esse.

Unless I'm translating this incorrectly, "se" seems to refer to
"ducem", which is in the accusative case. Or is it not a matter of the
exact case, but the grammatical role of the word, which in this case
acts as a subject for "polliceri"?

And what happens if I wanted to use the dative instead of the
accusative?

Any enlightenment would be appreciated.

Many thanks,
Adam

[Johannes Patruus]

Post by Adam
Hello all,
First just want to say that it's too bad there's so much crap being
posted to this newsgroup. Is there another, less cluttered newsgroup
where people are posting legitimate questions?

For discussion of all things Latin, including grammar, why not drop in on
<alt.language.latin>?

Johannes

Post by Adam
I have a question about the pronoun "se" and its use in a subordinate
subjunctive clause or indirect statement. According to my book (which
is really quite excellent!) unless modified by "ipse", occurances of
forms of "se" in a subordinate clause or indirect statement will refer
to the subject of the main clause. But how does that work if the main
clause is an impersonal such as "oportere"? One of the exercises in my
Pacis petendae causa ducem oportebat polliceri se nemini post bellum
nociturum esse.
Unless I'm translating this incorrectly, "se" seems to refer to
"ducem", which is in the accusative case. Or is it not a matter of the
exact case, but the grammatical role of the word, which in this case
acts as a subject for "polliceri"?
And what happens if I wanted to use the dative instead of the
accusative?
Any enlightenment would be appreciated.
Many thanks,
Adam

[Neeraj Mathur]

Post by Adam
I have a question about the pronoun "se" and its use in a subordinate
subjunctive clause or indirect statement. According to my book (which
is really quite excellent!)

Which book is that? I learned my Latin first from Wheelock, and loved it
while using it; it was only once I started my studies at Oxford that I
realised its deficiancies; I now find it generally best to have a few
grammars all at hand and then cross-reference them and synthesize a
consensus to find the answers to tougher problems...

Post by Adam
unless modified by "ipse", occurances of
forms of "se" in a subordinate clause or indirect statement will refer
to the subject of the main clause.

I would suggest that that's probably not the best maxim to learn, because
the usage of 'ipse' is probably not as consistent as your grammarian would
hope.

Kennedy (_Revised Latin Primer_) offers the following (in Section 316):

'The Reflexive Pronoun, Se, is used to refer:
(a) to the subject of the Simple Sentence or Subordinate Clause in which it
stands;
(b) to the subject of a Principle Sentence, if the Subordinate Clause in
which it stands represents something in the mind of that subject;
(c) to the subject of a Verb of Saying which introduces Indirect Speech.'

Although it seems I am suggesting replacing one maxim with another, I think
this one has a few advantages over the one you quoted. Your grammarian, I
think, is counting on the presence of ipse to apply in Kennedy's situation
(a) above; in other words, he is saying that if the subordinate clause does
not represent something in the mind of the subject of the principal clause,
he is hoping that will be signposted by an 'ipse'. Indeed, his own examples
may conform to that, but it is virtually certain that the usage of Latin
authors is not so easily predictable with regards to 'ipse'.

I would suggest that you cross-refernce with a check in the so-called
'Bradley's Arnold' - a fantastically detailed grammar (I think it's official
title is something boring like 'Latin Prose Composition'), and also in
Woodcock's _A New Latin Syntax_ who has several pages of discussion with
examples on the subject. The general consensus is to agree with this idea of
the subordinate clausing having to be something 'in the mind' of the subject
of the main clause if the se is to refer to the subject of the main clause;
otherwise, if the subordinate clause is not an idea 'in the mind' of the
subject of the main clause, the se referes to the subject of the subordinate
clause.

Sorry for this rather fluffy-sounding 'in the mind' thing, but if you look
over the pages of examples in Kennedy, Woodcock, Bradley's Arnold and some
of the other references (not sure what North and Hillard think here) then
you should eventually be able to see the pattern emerge.

Post by Adam
But how does that work if the main
clause is an impersonal such as "oportere"? One of the exercises in my
Pacis petendae causa ducem oportebat polliceri se nemini post bellum
nociturum esse.
Unless I'm translating this incorrectly, "se" seems to refer to
"ducem", which is in the accusative case.

Okay, the issue here is that there is a subordinate clause within a
subordinate clause. The 'se' belongs to the third, most nestled clause, and
refers to the subject of the second clause. Schematically:

1) oportebat
2) .......... ducem polliceri
3) .......... ............ se nemini nociturum esse

to be translated:
1) it was fitting
2) that the leader promise
3) that he would injure nobody.

The third clause is a 'promise' - therefore, it fulfils the 'in the mind'
criteria. (What the leader promises must be in the mind of the leader.)
Therefore, the 'se' refers to the subject of _its_ main clause, which in
this case is 'ducem polliceri'. The fact that this happens to be subordinate
to the impersonal 'oportebat' is irrelevant to the analysis. In all of our
rules above, we should have said that they go back _one level only_ of the
subordination; they do not refer to the absolute main clause.

You will see that 'se' is in the accusative case, and is modified by the
adjective (participle, rather) 'nociturum', which together form the subject
for the infinitive 'esse'. (While this explanation has a basic, formal
degree of correctness which I think makes it easier to see the relations,
you may prefer to see 'nociturum esse' as a kind of future infinitive, which
is how most grammarians describe it.)

Post by Adam
Or is it not a matter of the
exact case, but the grammatical role of the word, which in this case
acts as a subject for "polliceri"?

Exactly: case has nothing to do with it, you have to look for the subject of
the clause. (It is then relatively easy to see that in practise, you will
look for a word in the nominative if the verb happens to be finite and in
the accusative if it happens to be infinite, but focus on the meanings
rather than the shapes and work will be easier!)

Post by Adam
And what happens if I wanted to use the dative instead of the
accusative?

What do you mean here, exactly? The forms of 'se' must refer to a word which
functions as a subject for some clause - either the clause in which the
features, or the one above it. So, as a dative cannot usually be a subject,
a word in the dative cannot ordinarily be the antecedent of a form of 'se'.
Of course, se itself can be in any case for which it has forms, including
its dative 'sibi' - the case of 'se' is determined, like that of all
pronouns, by the clause in which it stands. The only exceptions to these
general rules are where you get 'sense' overriding form - things like
'relative attraction' or notional subjects (for examples, turn to Woodcock).

Here's one quick example, quoted by Kennedy, of 'sibi' in a subordinate
clause: 'impetrat a senatu ut dies sibi prorogaretur' (from Cicero) which
Kennedy translates as 'he prevailed upon the senate to postpone the date for
himself'. Here, 'sibi' refers to the subject of 'impetrat', and Kennedy
assumes that this means Cicero that that what was being 'impetrat'ed was in
the mind of the 'impetrat'er, if you see what I mean; hence, the sibi does
not refer to the subject of its own clause 'dies'.

Clear as mud? I hope that helped...

Neeraj Mathur

[Gordon]

In article <celg39$qbi$***@news.ox.ac.uk>, ***@chch.ox.ac.uk
says...

Post by Neeraj Mathur

Post by Adam
I have a question about the pronoun "se" and its use in a subordinate
subjunctive clause or indirect statement. According to my book (which
is really quite excellent!)

Which book is that? I learned my Latin first from Wheelock, and loved it
while using it; it was only once I started my studies at Oxford that I
realised its deficiancies; I now find it generally best to have a few
grammars all at hand and then cross-reference them and synthesize a
consensus to find the answers to tougher problems...

Post by Adam
unless modified by "ipse", occurances of
forms of "se" in a subordinate clause or indirect statement will refer
to the subject of the main clause.

I would suggest that that's probably not the best maxim to learn, because
the usage of 'ipse' is probably not as consistent as your grammarian would
hope.
(a) to the subject of the Simple Sentence or Subordinate Clause in which it
stands;
(b) to the subject of a Principle Sentence, if the Subordinate Clause in
which it stands represents something in the mind of that subject;
(c) to the subject of a Verb of Saying which introduces Indirect Speech.'
Although it seems I am suggesting replacing one maxim with another, I think
this one has a few advantages over the one you quoted. Your grammarian, I
think, is counting on the presence of ipse to apply in Kennedy's situation
(a) above; in other words, he is saying that if the subordinate clause does
not represent something in the mind of the subject of the principal clause,
he is hoping that will be signposted by an 'ipse'. Indeed, his own examples
may conform to that, but it is virtually certain that the usage of Latin
authors is not so easily predictable with regards to 'ipse'.
I would suggest that you cross-refernce with a check in the so-called
'Bradley's Arnold' - a fantastically detailed grammar (I think it's official
title is something boring like 'Latin Prose Composition'), and also in
Woodcock's _A New Latin Syntax_ who has several pages of discussion with
examples on the subject. The general consensus is to agree with this idea of
the subordinate clausing having to be something 'in the mind' of the subject
of the main clause if the se is to refer to the subject of the main clause;
otherwise, if the subordinate clause is not an idea 'in the mind' of the
subject of the main clause, the se referes to the subject of the subordinate
clause.
Sorry for this rather fluffy-sounding 'in the mind' thing, but if you look
over the pages of examples in Kennedy, Woodcock, Bradley's Arnold and some
of the other references (not sure what North and Hillard think here) then
you should eventually be able to see the pattern emerge.

Post by Adam
But how does that work if the main
clause is an impersonal such as "oportere"? One of the exercises in my
Pacis petendae causa ducem oportebat polliceri se nemini post bellum
nociturum esse.
Unless I'm translating this incorrectly, "se" seems to refer to
"ducem", which is in the accusative case.

Okay, the issue here is that there is a subordinate clause within a
subordinate clause. The 'se' belongs to the third, most nestled clause, and
1) oportebat
2) .......... ducem polliceri
3) .......... ............ se nemini nociturum esse
1) it was fitting
2) that the leader promise
3) that he would injure nobody.
The third clause is a 'promise' - therefore, it fulfils the 'in the mind'
criteria. (What the leader promises must be in the mind of the leader.)
Therefore, the 'se' refers to the subject of _its_ main clause, which in
this case is 'ducem polliceri'. The fact that this happens to be subordinate
to the impersonal 'oportebat' is irrelevant to the analysis. In all of our
rules above, we should have said that they go back _one level only_ of the
subordination; they do not refer to the absolute main clause.

Perhaps the shortest way out of this conundrum is to realize that what we
call impersonal verbs are usually verbs for which the subject is an
infinitive phrase. _ducem_polliceri_ is in fact the *subject* of
_oportebat_, and thus _se_ functions just as it should, taking as its
antecedent the subject of the main clause, or more precisely, the subject
of the subject of the main clause.

[tb+ (Thomas Bushnell, BSG)]

Post by Gordon
Perhaps the shortest way out of this conundrum is to realize that what we
call impersonal verbs are usually verbs for which the subject is an
infinitive phrase. _ducem_polliceri_ is in fact the *subject* of
_oportebat_, and thus _se_ functions just as it should, taking as its
antecedent the subject of the main clause, or more precisely, the subject
of the subject of the main clause.

Well, that's really not right in this case, but you can think of it
this way if you like. The actual grammar of oportebat is that it
takes a double object, one in the accusative, one as a complementary
infinitive.

[Neeraj Mathur]

Post by Adam
But how does that work if the main
clause is an impersonal such as "oportere"?

Hmm, it occurred to me that there's a more direct answer to this question.

First, Kennedy's 'mind-rule' reiterated:
in a subordinate clause, 'se' refers to the subject of the principal clause
if the sentence is in the mind of the subject of the principal clause; if
not, 'se' refers to the subject of the subordinate clause.

Now, about impersonal verbs:
the subject of an impersonal verb is a grammatical entity only, not a real
person; therefore, there is nothing which can be said to be 'in the mind of
the subject' of an impersonal verb.

The conclusion, a rather general one:
'Se' or any of its cases can never refer to the subject of an impersonal
verb in a clause dependent upon an impersonal verb. If an impersonal verb
introduces a dependent clause, 'se' must refer to the subject of the
dependent clause only.
_____

Now, your sentence had three levels and was thus a bit more complicated than
necessary. Here's a sentence with only two levels:

castris captis, oportet ducem se necare.

The main clause is 'oportet'; the secondary clause is 'ducem necare'. 'Se'
is in the secondary clause dependent on an impersonal verb; therefore, our
conclusion above applies, and 'se' refers to the subject of its own clause,
'ducem'. The fact that 'ducem' is accusative does not affect this (it is
forced on us because necare is an infinitive).
_____

Here is a trickier situation:

Cicero vult Caesaram se necare.

Here, we have a bit of difficulty: we can either interpret this as 'Cicero
wishes that Caesar kill Cicero' or as 'Cicero wishes that Caesar kill
Caesar'. If we apply the 'mind-rule', we might say that what Cicero wants is
in his mind, and therefore, 'se' refers to the subject of the main verb
'vult' and the first interpretation is correct. If we apply your
grammarian's 'ipse' rule, we see that there is no 'ipse' modifying 'se' and
so reach the same conclusion. But is this correct? Obivously, if I or some
other modern-day grammarian creates a sentence, we can create in
confirmation of whichever rule we pick; when trying to _read_ ancient Latin
authors, though, our rules (which are based on statistics or general
impressions) may not hold.

On this point, read Woodcock (sections 36 and 37). While he seems to agree
with the 'mind' thing (and, I think, would say it applies in this instance),
he says explicitly that many sentences are ambiguous, and he quotes various
sentences found in Roman authors that have this kind of ambiguity. He says,
essentially, that we must trust to context whenever we can, and hope that
the author would have provided enough to make it clear or picked an
alternate expression if it was impossible to make clear. He writes: 'Latin
authors do not, in fact, seem to have been much worried by the ambiguity, if
the sense was clear from context'.

Here's one of his examples (Caesar, B.G. 1.36.6): 'Ariovistus respondit
neminem secum sine sua pernicie contendisse'. The two reflexives, secum and
sua, both refer to separate things and could refer to either one; he
translates it, using context as a guide, 'Ariovistus replied that no one had
opposed him (Ariovistus) without bringing about his own downfall (ie, not
Ariovistus' downfall).'

He goes on to say this: 'The statement sometimes made, that the ambiguity is
avoided by the use of 'ipsum' for the indirect reflexive, is erroneous, at
least for classical Latin.' He has lots of examples, with discussion, to
explain his views. Do take a look at this book and read it carefully.

Sorry, I thought I wouldn't ramble in this post!

Neeraj Mathur

[tb+ (Thomas Bushnell, BSG)]

Post by Neeraj Mathur
What do you mean here, exactly? The forms of 'se' must refer to a word which
functions as a subject for some clause - either the clause in which the
features, or the one above it. So, as a dative cannot usually be a subject,
a word in the dative cannot ordinarily be the antecedent of a form of 'se'.

In the middle ages "se" starts to get used even more flexibly than
that, but the sample sentence here demonstrates the shift in process.

Strictly speaking, in "oportebat ducem pollicari", the clause "ducem
pollicari" is not really indirect speech; ducem is really the
accusative object of "oportebat", but it is starting to feel like the
accusative subject of "pollicari". But actually, it's not; oportebat
takes two objects, a person, and a complementary infinitive.

So this is already an example where "se" is referring to something
that is not really the subject of any clause at all. But in the
*meaning*, even if not the strict grammar, it certainly is.

Thomas

[Neeraj Mathur]

Post by tb+ (Thomas Bushnell, BSG)
In the middle ages "se" starts to get used even more flexibly than
that, but the sample sentence here demonstrates the shift in process.
Strictly speaking, in "oportebat ducem pollicari", the clause "ducem
pollicari" is not really indirect speech; ducem is really the
accusative object of "oportebat", but it is starting to feel like the
accusative subject of "pollicari". But actually, it's not; oportebat
takes two objects, a person, and a complementary infinitive.
So this is already an example where "se" is referring to something
that is not really the subject of any clause at all.

Ah, I see what you're saying. However, I think that the development in
question here is not that of 'se' referring to non-subject nouns, but rather
to something much earlier: the development of the acc + inf construction.
So, what was originally a double infinitive construction was in time
reanalysed by the Latins (or their ancestors) as an acc + inf clause; once
this reanalysis is made and ducem is thought of, not as one of two
accusative objects but as an oratio obliqua subject of an infinitive, the
representation of it by se follows naturally.

So I see the grammatical development as one of reanalysis of a double
infinitive as an acc + inf, rather than being one of the expanding use of se
to refer to non-subject nouns. To confirm or deny, however, we would need to
amass a chronological pool of examples.

To Adam: where did your sentence come from? Was it a quotation from an
author (if so, which one) or was it invented by the author of the textbook?

Neeraj

[tb+ (Thomas Bushnell, BSG)]

Post by Neeraj Mathur
Ah, I see what you're saying. However, I think that the development in
question here is not that of 'se' referring to non-subject nouns, but rather
to something much earlier: the development of the acc + inf construction.

Well, certainly both are going on as well. The acc + inf construction
is a near-universal in Indo European languages. For example, in a
Germanic language familiar to us all, we have: "I want him to tell me
the truth." In that sentence, what I want is not "him", but "that he
tell me the truth", making it clear that "him" is not a direct
object.

It might seem that "oportebat" is the same, but actually no, indeed, I
believe the history is that the infinitive is the subject of
oportebat, and the accusative is the *object*, so that it is something
like: "To promise to XXX suits the leader." (Except that "suits" is
not a good translation of "oportebat", of course.)

So I'm skeptical about the idea that the acc + inf is a relative
latecomer to Latin. I think it's really just a very old thing, which
pops up frequently in different IE languages for different things.

Post by Neeraj Mathur
So, what was originally a double infinitive construction was in time
reanalysed by the Latins (or their ancestors) as an acc + inf clause; once
this reanalysis is made and ducem is thought of, not as one of two
accusative objects but as an oratio obliqua subject of an infinitive, the
representation of it by se follows naturally.

Now *this* I agree with. The acc + inf construction already exists,
and other things get reinterpreted as it.

Post by Neeraj Mathur
So I see the grammatical development as one of reanalysis of a double
infinitive as an acc + inf, rather than being one of the expanding use of se
to refer to non-subject nouns. To confirm or deny, however, we would need to
amass a chronological pool of examples.

Well, the reason I phrased it as I did was probably because I'm so
used to medieval latin. By that time, expansion of acc + inf has
ceased and the trend is to use "quot" and "ut" clauses instead (you
see "oportebat ut ..." even classically). And, the use of "se"
becomes even looser, so that it becomes about as loose as "suus"
always was.

Thomas

[Neeraj Mathur]

Post by tb+ (Thomas Bushnell, BSG)
It might seem that "oportebat" is the same, but actually no, indeed, I
believe the history is that the infinitive is the subject of
oportebat, and the accusative is the *object*, so that it is something
like: "To promise to XXX suits the leader." (Except that "suits" is
not a good translation of "oportebat", of course.)
So I'm skeptical about the idea that the acc + inf is a relative
latecomer to Latin. I think it's really just a very old thing, which
pops up frequently in different IE languages for different things.

Right - as you realised below, I wasn't talking about inventing the
construction, merely about extending it. Your idea that the infinitive was
originally the subject of the verb, which then took an accusative direct
object, seems to be the history that Woodcock prefers for this verb as well.

That said, I have many reservations about attributing the acc + inf
construction to Indo-European, not least because 'infinitive' as a
morphological category is a meaningless, or at best anachronistic, term
applied to the parent language. Indo-European had many verbal nouns, which
seem to have taken case inflections just like other nouns; many of these
survive in Greek (hence the plethora of inf endings there, with due credit
given to analogical extension), and there are three in Latin, two also
survived into Anglo-Saxon. The difficulty of trying to arrive at some
acceptable IE syntax, though, can be highlighted if we consider that (many
of) the grammatical functions of the Latin inf are carried in Sanskrit by a
form that is morphologically equivalent to the Latin acc supine.

Further, comparative evidence of the syntax seems a bit suspect to me.
Sanskrit has no equivalent to the acc + inf; the Greek equivalent is much
more complex (since it can be a nom + inf if its subject is the same as the
sentence); the Germanic evidence is less firm than you suggest. There is no
acc + inf in Old Norse, and in Old English there is both an acc + inf and
just an infinitive without a noun subject, with the subject 'understood'.
There is some debate as to whether this is a passive infinitive; in light of
the major impact of Latin on the syntax of OE I wonder if this is the older
construction, expanded under the influence of the Latin acc + inf (I don't
have the data to support or deny this, though). Is there anybody here who
can shed light on any of the other branches? I would be most interested to
see what Hittite has to say.

Anyway, discussions of IE syntax cannot be anything but primitive, I think,
since there simply isn't a way to extend the rigourous comparative method
that works so well with phonology and morphology to syntax. If you compare
the Romance languages, for instance, I'm not sure how you could derive a
Latin acc + inf; the verbal systems of the modern Indian languages differ
hugely from Sanskrit, but have an odd similarity to modern Farsi and the
modern Celtic tongues - parallel development is unmistakably at play here,
and nothing can be back-computed from it.

For all these reasons, I prefer to consider the acc + inf as essentially a
Latin (or at least Italic) syntactical phenomenon; taking it back further
takes us beyond the realm of evidence. It may have existed there, but this
cannot be proven. (Woodcock himself has an interesting theory on the
development of the Latin infinitive which is based on the fact that it is
formed from a local case; I'm not sure how he reconciles this with his view
of 'oportet' but I suppose with a bit of chronological play it could be
done; in this case Latin infinitive constructions would have emerged fairly
early into the history of Italic, when the case forms of IE in the verbal
noun were still recognisable.)

Post by tb+ (Thomas Bushnell, BSG)
Well, the reason I phrased it as I did was probably because I'm so
used to medieval latin. By that time, expansion of acc + inf has
ceased and the trend is to use "quot" and "ut" clauses instead (you
see "oportebat ut ..." even classically).

Right, this must be the origin of Romance constructions (Fr / Sp 'que' etc).
Would you know any good books on the emergence of Romance from Vulgar Latin?
Did Ecclesiastic Latin still influence Romance development?

What is your research on, in the medieval world?

Neeraj Mathur

[tb+ (Thomas Bushnell, BSG)]

Post by Neeraj Mathur
That said, I have many reservations about attributing the acc + inf
construction to Indo-European, not least because 'infinitive' as a
morphological category is a meaningless, or at best anachronistic, term
applied to the parent language.

Ah, my usage was confusing. When I mean the parent language, I was
taught to say "proto-Indo-European"; by "Indo-European" I mean
languages which are extant. I'm grateful for your descriptions of the
complexities of the latter: very helpful to me.

Thomas

[Neeraj Mathur]

Post by tb+ (Thomas Bushnell, BSG)
Ah, my usage was confusing. When I mean the parent language, I was
taught to say "proto-Indo-European"; by "Indo-European" I mean
languages which are extant.

Okay, I see; the consensus opinion of my lecturers and the Professor of
Philology here has been to use 'Indo-European' for the parent language
immediately before it split, and 'Proto-Indo-European' for older, internally
reconstructed forms of the parent language (so they would say that the IE
word for 'father' int he nom. sg. was 'pHter' (long e), but that in Proto IE
it was 'pHters' (short e)). They say 'daughter languages' or 'attested
languages' or sometimes 'IE languages' for the extant ones.

Post by tb+ (Thomas Bushnell, BSG)
I'm grateful for your descriptions of the
complexities of the latter: very helpful to me.

Way cool - if I've been useful, excellent! :-) Happy to help, if my
ramblings ever do.

Neeraj Mathur

[tb+ (Thomas Bushnell, BSG)]

Post by Neeraj Mathur
Right, this must be the origin of Romance constructions (Fr / Sp 'que' etc).

True without a doubt.

Post by Neeraj Mathur
Would you know any good books on the emergence of Romance from
Vulgar Latin?

There are bunches; I'll poke around and see if I can get together a
few recommendations.

Post by Neeraj Mathur
Did Ecclesiastic Latin still influence Romance development?

It's extremely hard to tell. In countries that spoke Romance, there
was not a conscious distinction between Latin and vernacular until at
least AD 800. We find people apparently treating written Latin as
just very peculiar spelling rules. At the same time, the state of
learning was so terribly poor; Merovingian Latin is just outrageously
awful. One of the things that the Carolingian reform did was to
require that the liturgy would be celebrated "as it's written", that
is, the words would pronounced as they are written. That indicates
that earlier, they were pronounced as vernacular words (more or less,
maybe), and that by AD 800 the divergence had become sufficiently
sharp as to need a remedy.

So one problem is that people might be speaking Romance, and when they
write down "what we're speaking" they write down Latin! (Notice how
different the words of this post are written down from the actual
pronunciation...) So this makes it even harder to track.

However, there is a pretty substantial break in France in this period,
because of the total breakdown of education under the Merovingians.
And your question is so very hard to answer because there is such a
great lack of documentary evidence on both sides of the question.

Post by Neeraj Mathur
What is your research on, in the medieval world?

I study 12th century philosophy, specifically, the moral philosophy of
Peter Abelard.

Thomas

[Rich Alderson]

At the same time, the state of learning was so terribly poor; Merovingian
Latin is just outrageously awful. One of the things that the Carolingian
reform did was to require that the liturgy would be celebrated "as it's
written", that is, the words would pronounced as they are written. That
indicates that earlier, they were pronounced as vernacular words (more or
less, maybe), and that by AD 800 the divergence had become sufficiently sharp
as to need a remedy.

It also made for a requirement that the clergy had to be literate, rather than
learning the Mass by rote from others: No more "In nomine Patris, et Filiae,
et Spiritu Sancto" problems.

[John Briggs]

Post by tb+ (Thomas Bushnell, BSG)
It's extremely hard to tell. In countries that spoke Romance, there
was not a conscious distinction between Latin and vernacular until at
least AD 800. We find people apparently treating written Latin as
just very peculiar spelling rules. At the same time, the state of
learning was so terribly poor; Merovingian Latin is just outrageously
awful. One of the things that the Carolingian reform did was to
require that the liturgy would be celebrated "as it's written", that
is, the words would pronounced as they are written. That indicates
that earlier, they were pronounced as vernacular words (more or less,
maybe), and that by AD 800 the divergence had become sufficiently
sharp as to need a remedy.

There's something wrong somewhere with this - Ecclesiastical (and to a large
extent Classical, but that's a different can of worms) Latin words were
pronounced as if they were vernacular words until about 1900 - in Italy they
still are, of course. Do you mean that the Latin words were pronounced the
same way as differently-spelled vernacular words? Or that vernacular words
were used instead? (Is that different?) Or are you really saying that the
liturgy was celebrated in the vernacular until AD 800?

[tb+ (Thomas Bushnell, BSG)]

Post by John Briggs
There's something wrong somewhere with this - Ecclesiastical (and to a large
extent Classical, but that's a different can of worms) Latin words were
pronounced as if they were vernacular words until about 1900 - in Italy they
still are, of course. Do you mean that the Latin words were pronounced the
same way as differently-spelled vernacular words? Or that vernacular words
were used instead? (Is that different?) Or are you really saying that the
liturgy was celebrated in the vernacular until AD 800?

I mean that until the Carolingian reform, in France at least, people
talked as if written Latin were the same thing as written
Romance--there WAS NO written Romance, actually. The people spoke
Romance, and they wrote down Latin words (usually badly). They
themselves usually called their own language alternately Latin and
Roman, with no real difference in meaning between the two terms.

By and large, they did not consciously notice than their language had
diverged massively from the liturgy they were reading. And they
didn't think "these are weird spelling rules, unlike French", because
there were no French spelling rules. There was just the one language,
written in the one way. And you knew how to write it, because you
could see the old texts.

It is only with the scholarship of someone like Alcuin--who came from
a Germanic country, note!--which exposed the truth, and people for the
first time became conscious that the vernacular was very different
from Latin.

So the liturgy was, sort of, celebrated in the vernacular until AD
800, sort of. It's perhaps something like reading Chaucer and
pretending that all the words are current words, and sliding over the
weird spelling. You might read:

"When that April with his showers sweet, the drought of March has
pierced to the root, and bathed every vein in such liquor, of which
virtue engendered is the flower; when Zephyrus eek with his sweet
breath, inspired has in every holt and heath, the tender crops and the
young sun has in the Ram his half course run, and small fouls make
melody, that sleep all the night with open eye, (So pricks them Nature
in their courages), then long folk to go on pilgramages, and palmers
for to seek strange strands, to foreign hallows, known in sundry
lands; and especially from every shire's end of England to Canterbury
they went, the holy blissful martyr for to seek, that he has helped
whom that they were seek."

It's like all the words are ones you know (but some archaic, like
"eek", and others with weird usage, like "courages" and "foreign", and
still others that sound poetical but were really just ordinary speech,
like "liquor"), and all the grammar makes sense but feels old. If you
were used to it, you could easily just think that there was no big
difference between you and Chaucer in the language you spoke.

Then along comes Alcuin. He explains to you how to pronounce Middle
English. He explains the way the grammar really works, and shows the
constructions that mean something quite different from what the same
words now mean; he points out to you the way the noun inflections
work (before when writing you more or less just guessed--which is part
of why Merovingian Latin is so terrible); and he says "now you have to
pronounce all the syllables, and the vowels are different from the way
you are saying them, and you're getting the accents all wrong".

You *thought* that Chaucer and you had the same language, but it turns
out you were wrong.

Thomas

[Neeraj Mathur]

Post by tb+ (Thomas Bushnell, BSG)

Post by Neeraj Mathur
Would you know any good books on the emergence of Romance from Vulgar Latin?

There are bunches; I'll poke around and see if I can get together a
few recommendations.

Thank you, I'd appreciate that.

Post by tb+ (Thomas Bushnell, BSG)

Post by Neeraj Mathur
Did Ecclesiastic Latin still influence Romance development?

<snip reply>

Thanks for this, which is quite informative, despite the limitations of the
question. I've always been a bit unclear as to how what happened, but your
examples (and your Chaucer demo) have been very useful. I do wonder if it's
possible to estimate how strong a force literacy is in arresting language
change - for instance, I've often thought that Modern English simply will
not evolve in any significant way in teh directions it is headed, because of
the extraordinary uber-literacy of Anglophones (comparatively) which I don't
see diminishing, political/social disasters discounted. (For instance, I
should be qutie pleased to see English's descendants settle down into decent
head-first or tail-first languages, rather than the awful mish-mash we have
now!)

Post by tb+ (Thomas Bushnell, BSG)

Post by Neeraj Mathur
What is your research on, in the medieval world?

I study 12th century philosophy, specifically, the moral philosophy of
Peter Abelard.

Ah, I see. Here's another rough(ish) question: what makes Descartes such a
revolutionary thinker, when his writings (chiefly the Meditations) cannot be
understood outside of a Mediaeval context? (At least, I had to learn some
basics of Mediaeval logic etc. before I could get some of his arguments!)

Neeraj

[tb+ (Thomas Bushnell, BSG)]

Post by Neeraj Mathur
Thanks for this, which is quite informative, despite the limitations of the
question. I've always been a bit unclear as to how what happened, but your
examples (and your Chaucer demo) have been very useful. I do wonder if it's
possible to estimate how strong a force literacy is in arresting language
change - for instance, I've often thought that Modern English simply will
not evolve in any significant way in teh directions it is headed, because of
the extraordinary uber-literacy of Anglophones (comparatively) which I don't
see diminishing, political/social disasters discounted. (For instance, I
should be qutie pleased to see English's descendants settle down into decent
head-first or tail-first languages, rather than the awful mish-mash we have
now!)

Ah, there have been lots of ways that English is changing!

For example, the growth of "you all" as a second person plural
pronoun. In the southern US, you have the flexibility of *three*
second person pronouns: you, y'all, all y'all. "Y'all" is just simple
plural, which has lost the "all" sense so much that "all y'all" means
"all of you"--that is, the whole group in earshot.

Outside the south, we cringe at "y'all"--but we do say "you all". If
you listen, you'll hear lots of pronunciation shifts going on.
Another change that I wholeheartedly support is the use of "they" as a
gender-neutral third person singular pronoun. This has become so
common in America that it no longer gets blinked at, though it is
somewhat restricted; it can only be used where there is some other
clearly singular antecedent sitting around. I expect that this
restriction on its usage will vanish over time.

Literacy notwithstanding, people learn their languagy orally, and they
speak the language not of their parents, but of their childhood
peers. And then they learn to read and write, and they learn to do so
by rules. We are so terribly used to English writing having only a
loose phonological shift, that we don't notice the ways the language
is changing.

For example, past tenses in -ed used to have a separate syllable. So
the second word of "He walked to the store" had two syllables. Then
in recent history--the last few hundred years--this got syncopated.
But people *remembered* the old pronunciation, so they wrote
"walk'd". But have have *forgotten* the old pronunciation, so we
write "walked" and just think of it as a weird spelling rule, rather
than what it truly is: we are spelling the way the word *used* to be
pronounced.

Post by Neeraj Mathur

Post by tb+ (Thomas Bushnell, BSG)

Post by Neeraj Mathur
What is your research on, in the medieval world?

I study 12th century philosophy, specifically, the moral philosophy of
Peter Abelard.

Ah, I see. Here's another rough(ish) question: what makes Descartes such a
revolutionary thinker, when his writings (chiefly the Meditations) cannot be
understood outside of a Mediaeval context? (At least, I had to learn some
basics of Mediaeval logic etc. before I could get some of his arguments!)

Lol. You are asking the *wrong* person! I have almost no patience
for the supposed "first modern philosopher" status of Descartes. He
did many revolutionary things--but he was part of a continuous
stream. However, he so earnestly engaged in a rather self-centered
campaign of pronouncing on his own importance and revolutionary
character, that people bought it.

Yeah, there's lots new in Descartes, but the myth that he left behind
the middle ages is just so much self-projection.

Thomas

[John W. Kennedy]

Post by tb+ (Thomas Bushnell, BSG)
It's extremely hard to tell. In countries that spoke Romance, there
was not a conscious distinction between Latin and vernacular until at
least AD 800.

And even 500 years later Dante is still somewhat hesitant to regard
Italian as a "language" in the same sense that Latin is a "language",
even as he is championing it.

[John Briggs]

Post by John W. Kennedy

Post by tb+ (Thomas Bushnell, BSG)
It's extremely hard to tell. In countries that spoke Romance, there
was not a conscious distinction between Latin and vernacular until at
least AD 800.

And even 500 years later Dante is still somewhat hesitant to regard
Italian as a "language" in the same sense that Latin is a "language",
even as he is championing it.

That sounds a bit like my old Latin master claiming that all Italians could
understand Latin perfectly well. He said that in North Africa during WWII
he only had to point his revolver at an Italian prisoner and give
instructions in Latin for them to be carried out :-)

[tb+ (Thomas Bushnell, BSG)]

Post by John Briggs
That sounds a bit like my old Latin master claiming that all Italians could
understand Latin perfectly well. He said that in North Africa during WWII
he only had to point his revolver at an Italian prisoner and give
instructions in Latin for them to be carried out :-)

Hard to imagine.... surely he only knew the British Received
Pronunciation, and the Italian would only understand Church Latin? :)

[Adam]

I see, so what you are suggesting is that it's not the leader doing
the promising, but someone else promising the leader, because the
subject of oportebat is the infinitive, polliceri.

It says in my book (again, Latin: An Intensive Course from the
University of California Press) that oportet may take an object which
is the person or thing concerned. It goes on to talk about licet,
necesse est (which can take a dative or an ut clause) as well as
piget, pudet, taedet, miseret, and paenitet, all of which take an
accusative of the thing or person concerned and a genitive of the
person or thing causing the emotion. This last group of impersonals
may employ an infinitive, quod clause of fact, or neuter demonstrative
pronoun as the source of the emotion, in which case they behave as the
subject. (This book only addresses classical Latin of Cicero's time.)

Thus, I'm gathering that oportet always takes an accusative as the
person or thing concerned and an infinitive to complete the sense, in
the same way as: "I want to sleep", and not as a subject (which the
book explicitly allows for piget, pudet, taedet, miseret, and
paenitet, and also for refert and interest). But I don't have the book
with me at the moment so I can't double check if I'm actually telling
you what the book says, or just how I remember it.

It's interesting, by the way, hearing you talk about the expansion of
the use of "se", and someone even mentioned how loose the use of suus
was. In fact, when I first started with Latin I found the use of suus
confusing because in French, son/sa are simply the third-person
possessive pronouns.

Post by tb+ (Thomas Bushnell, BSG)

Post by Neeraj Mathur
Ah, I see what you're saying. However, I think that the development in
question here is not that of 'se' referring to non-subject nouns, but rather
to something much earlier: the development of the acc + inf construction.

Well, certainly both are going on as well. The acc + inf construction
is a near-universal in Indo European languages. For example, in a
Germanic language familiar to us all, we have: "I want him to tell me
the truth." In that sentence, what I want is not "him", but "that he
tell me the truth", making it clear that "him" is not a direct
object.
It might seem that "oportebat" is the same, but actually no, indeed, I
believe the history is that the infinitive is the subject of
oportebat, and the accusative is the *object*, so that it is something
like: "To promise to XXX suits the leader." (Except that "suits" is
not a good translation of "oportebat", of course.)
So I'm skeptical about the idea that the acc + inf is a relative
latecomer to Latin. I think it's really just a very old thing, which
pops up frequently in different IE languages for different things.

Post by Neeraj Mathur
So, what was originally a double infinitive construction was in time
reanalysed by the Latins (or their ancestors) as an acc + inf clause; once
this reanalysis is made and ducem is thought of, not as one of two
accusative objects but as an oratio obliqua subject of an infinitive, the
representation of it by se follows naturally.

Now *this* I agree with. The acc + inf construction already exists,
and other things get reinterpreted as it.

Post by Neeraj Mathur
So I see the grammatical development as one of reanalysis of a double
infinitive as an acc + inf, rather than being one of the expanding use of se
to refer to non-subject nouns. To confirm or deny, however, we would need to
amass a chronological pool of examples.

Well, the reason I phrased it as I did was probably because I'm so
used to medieval latin. By that time, expansion of acc + inf has
ceased and the trend is to use "quot" and "ut" clauses instead (you
see "oportebat ut ..." even classically). And, the use of "se"
becomes even looser, so that it becomes about as loose as "suus"
always was.
Thomas

[tb+ (Thomas Bushnell, BSG)]

Post by Adam
I see, so what you are suggesting is that it's not the leader doing
the promising, but someone else promising the leader, because the
subject of oportebat is the infinitive, polliceri.

No. The leader is promising. It is fitting that the leader promise
to ...

[Rolleston]

Post by Neeraj Mathur
Okay, the issue here is that there is a subordinate clause within a
subordinate clause. The 'se' belongs to the third, most nestled clause, and
1) oportebat
2) .......... ducem polliceri
3) .......... ............ se nemini nociturum esse

Are there good reasons for regarding (2) and (3) as clauses?

Allen & Greenough's New (now Old!) Latin Grammar has this:

The Infinitive with Subject Accusative is not strictly a clause,
but in Latin it has undergone so extensive a development
that it may be so classed. [Sec. 562]

The terminology itself may not be that important, as long as
we can be sure which grammatical structures you refer to.

Post by Neeraj Mathur
What do you mean here, exactly? The forms of 'se' must refer to a word which
functions as a subject for some clause - either the clause in which the
features, or the one above it.

When you write "functions as subject", are you referring to the
grammatical subject? Lewis and Short give some examples in
which "se" refers to a logical subject rather than a grammatical
subject (see I.B.1 in http://tinyurl.com/26sqt).

Examples:

neque praeter se umquam ei servos fuit
Plautus Capt. 3.4.48

ut quam minimum spatii ad se colligendos Romanis daretur
Caesar B. G. 3.19

Gallica acies nullum spatium respirandi recipiendique se dedit
Livy AUC 10.28.11

proelium cum fiducia sui commissum est
Livy AUC 7.33.5

But let's not over-emphasize the importance of individual examples.

R.

[Neeraj Mathur]

Post by Rolleston
Are there good reasons for regarding (2) and (3) as clauses?
The Infinitive with Subject Accusative is not strictly a clause,
but in Latin it has undergone so extensive a development
that it may be so classed. [Sec. 562]
The terminology itself may not be that important, as long as
we can be sure which grammatical structures you refer to.

What definition, exactly, do Allen and Greenough seem to use for 'clause' -
do they mean only something with a finite verb? If so, then they are not
clauses; then we would rework our 'se' rules to say 'subject of a finite
clause or of an infinitive'. In any case, I want to regard these as two acc
+ inf units (structures? groups of words? if not clause, what would we call
them?) and to then say that 'se' can refer to this accusative, because it is
a subject of an infinitive. So the terminology is not that important; I
think calling it a clause, as A & G permit it to be classed, makes our
'rules' tidier.

Woodcock calls it a 'noun-phrase' and writes this (Section 25): "The next
step in syntactical development was an associational shift, whereby the
infinitive with a verb which already had an object was dissociated from the
finite berb and associated with the accusative object, so that the word in
the accusative together with the infinitive formed a noun-phrase in which
the accusative appeared to be the subject of the infinitive."

So here again, the important point is that the accusative functions as a
subject, which qualifies it to be referred to by 'se'.

Post by Rolleston
When you write "functions as subject", are you referring to the
grammatical subject?

Yes.

Post by Rolleston
Lewis and Short give some examples in
which "se" refers to a logical subject rather than a grammatical
subject (see I.B.1 in http://tinyurl.com/26sqt).
But let's not over-emphasize the importance of individual examples.

Yes that's true - all of my books give similar examples. I vaguely suggested
to the original questioner that, in such cases, he should trust to the
sense, context and his instinct in working out the antecedent of 'se'; I
think I tried not to talk about them to avoid complicating the issue for a
learner.

I also suggested that these sorts of situations be studiously avoided when
composing in Latin, because (as it seems to me) this particular idiom was
based on a nuance that, while fully natural to native Latin usage, is likely
to be misapplied by foreigners two millenia later, at least until they
develop several years of reading experience. I think that the law as far as
prose comp is concerned should allow 'se' to refer only to subjects (nom of
finite verbs, acc of infinitives), while all the same an awareness is
maintained of the fact that actual writers used it in a slightly less
predictable fashion.

Neeraj

[tb+ (Thomas Bushnell, BSG)]

Post by Adam
Pacis petendae causa ducem oportebat polliceri se nemini post bellum
nociturum esse.
Unless I'm translating this incorrectly, "se" seems to refer to
"ducem", which is in the accusative case. Or is it not a matter of the
exact case, but the grammatical role of the word, which in this case
acts as a subject for "polliceri"?

"se" refers to "ducem" in that sentence. But "ducem," while in the
accusative case, is really the subject of "polliceri" in the sense of
the passage, if not the literal meaning.

Post by Adam
And what happens if I wanted to use the dative instead of the
accusative?

"sibi"

[Adam]

there have been a lot of responses and i have to admit, it will take a
bit to digest. a few things i neglected to point out beforehand:

1. the sentence used came from the exercises of my book and is not
"authentic" latin in that no latin native speaker wrote it (the book
is Moreland and Fletcher, Latin: An Intensive Course from University
of California Press), and
2. this book is not a grammarian in the sense that it is a compendium
of latin grammar, but is designed to be used to teach latin; concepts
are introduced progressively, and so the discussion of the reflexive
is treated in separate places; i just mentioned the one that i thought
was relevant and was asking for confirmation or refutation of my
thinking

it is true what you point out, thomas, that ducem is the accusative of
the verb oportere. but the reflexive aside, the construction still
seems to me to be subject-accusative, because the way i'm translating
the sentence (and perhaps erroneously) is that it's the leader who is
doing the promising. is the subject-accusative contruction only
reserved for indirect statement? anyway, to consider something
necessary or proper involves thought.

so although i still have to read through all these posts, i'm still
confused about whether 'se' can refer to an subject in the accusative,
or whether the sentence is in fact not a proper latin sentence.

as far as the dative case, i'll put it this way: what if the sentence
started thus:

Pacis petendae causa duci oportebat polliceri ....

many thanks!

adam

Post by tb+ (Thomas Bushnell, BSG)

Post by Adam
Pacis petendae causa ducem oportebat polliceri se nemini post bellum
nociturum esse.
Unless I'm translating this incorrectly, "se" seems to refer to
"ducem", which is in the accusative case. Or is it not a matter of the
exact case, but the grammatical role of the word, which in this case
acts as a subject for "polliceri"?

"se" refers to "ducem" in that sentence. But "ducem," while in the
accusative case, is really the subject of "polliceri" in the sense of
the passage, if not the literal meaning.

Post by Adam
And what happens if I wanted to use the dative instead of the
accusative?

"sibi"

[tb+ (Thomas Bushnell, BSG)]

Post by Adam
it is true what you point out, thomas, that ducem is the accusative of
the verb oportere. but the reflexive aside, the construction still
seems to me to be subject-accusative, because the way i'm translating
the sentence (and perhaps erroneously) is that it's the leader who is
doing the promising.

Yes, that's the correct meaning. And, the upshot of the discussion
here is that a classical author would have perceived it that way too.

Post by Adam
is the subject-accusative contruction only reserved for indirect
statement?

Yes, if you like, but you need to take an expansive view of indirect
statement! Anyhow, the best rough translation of an
accusative-infinitive construction is as an English "that" clause;
doing such will almost always get the meaning. In this case, "it is
fitting that the leader promise ..."

Post by Adam
anyway, to consider something necessary or proper involves thought.

Yes, but "oportet" doesn't imply that someone *considers* it
necessary: "oportet" here doesn't mean that the leader himself thinks
it's the right thing to do, but that the speaker does.

Post by Adam
so although i still have to read through all these posts, i'm still
confused about whether 'se' can refer to an subject in the accusative,
or whether the sentence is in fact not a proper latin sentence.

"se" refers to the subject. When one sentence is embedded in another,
it tends (but only *tends*) to "reach out" of that sentence to the
subject of the outer sentence (as here). There is no rigid rule.

Thomas

[Neeraj Mathur]

Post by Adam
so although i still have to read through all these posts, i'm still
confused about whether 'se' can refer to an subject in the accusative,
or whether the sentence is in fact not a proper latin sentence.

Yes it can:
1) Subjects of an infinitive verb are in the accusative case;
2) 'se' can refer to *any* subject, including one in the accusative. This is
'proper' Latin. That's why I recommended that you parse the sentence in the
way i suggested - go over my earlier posts, and it should get clear.

Post by Adam
as far as the dative case, i'll put it this way: what if the sentence
Pacis petendae causa duci oportebat polliceri ....

This is *not* proper Latin - a dative cannot be used in this way. (In odd
situations, 'se' has been used to refer to non-subject nouns; here, you
would trust to the sense of the sentence and go on your instincts. When you
are creating a Latin sentence, you should never do this - use se to refer
only to the subject of a verb, which will be nominative with a finite verb,
and accusative with an infinitive.)

Neeraj Mathur