[B-Greek] accentuation of proclitic and enclitic lemmata

James Tauber jtauber at jtauber.com
Sun Aug 28 23:19:51 EDT 2005


On 28/08/2005, at 11:42 PM, Carl W. Conrad wrote:

>
> On Aug 28, 2005, at 4:55 AM, James Tauber wrote:
>
>
>> The next release of MorphGNT will be a major release that includes an
>> extra column: the surface form of the word *without* contextual
>> changes like elision, movable-nu and various accentual changes. This
>> new context-free surface form is what you might expect in a paradigm
>> listing or if you asked someone to provide a particular inflected  
>> form.
>>
>
> This is less than altogether clear to me. Perhaps an example or two  
> would help.

- For KAI/ the new column would have KAI/ even though the text itself  
might have KAI\
- For A)LLA/ the new column would have A)LLA/ even though the text  
itself might have A)LL'
- (answering your question below) For POIE/W, the new column would  
have POIEI=N or E)POIHSA/MEN or E)POI/HSE(N) or whatever the  
inflected form is (but notice the movable-nu is included in parentheses)
- EI)MI/ in the 3rd person plural active indicative would have EI)SI/ 
(N) even though the text might have EI)SI\N
- the accusative singular of GUNH/ would have GUNAI=KA in the new  
column in Mt 1.20 even though the text column has GUNAI=KA/ because  
of the following enclitic SOU

So while the text may have a grave accent or two accents or an  
apostrophe marking elision, the new column never will.

The real heart of the question I'm asking here is:

a) what are the exceptions to the notion that the new column should  
always have one (non-grave) accent?

Certainly proclitics are a reasonable candidate. Foreign words that  
are unaccented in the UBS text are another candidate. The real  
question is the enclitics and here only the singular oblique personal  
pronouns and the indefinite pronoun are at issue.

b) what are the exceptions to the notion that the lemma column should  
always have one (non-grave) accent?

Again, the proclitics are a reasonable candidate. But what about  
enclitics? All enclitics I have encountered have accents in their  
dictionary form. Only the indefinite TIS might be controversial but  
both BDAG and Smyth still choose to accent this, although with a  
grave to distinguish from the interrogative.

> Although you're speaking below chiefly of forms of enclitics, it  
> seems to me that you are NOT speaking of the lemmata, some of which  
> are problematic in themselves. For instance the lemma of POIW IS  
> POIEW, even though that form never appears (unless it be in an  
> Ionic dialect form that doesn't contract vowels, as Herodotus);  
> will your context-free form for POIEW be POIE-? Or am I  
> misunderstanding completely what you're talking about when you  
> write "what you might expect in a paradigm listing or if you asked  
> someone to provide a particular inflected form"? For a noun like  
> ANHR or like QUGATHR would this column offer ANDR- or QUGATR-?

Eventually I want to include something like this but for now, I'm  
going the other way - e.g. the dative plural of ANHR.

>
>> A few of issues have arisen around the proper presentation of accents
>> on proclitics and enclitics in the fields in my database *other than*
>> the text itself. In other words, the context-free surface form and
>> the lemma.
>>
>> 1. I assume that it is always proper for both the lemmata and  
>> context-
>> free surface form of proclitics to appear unaccented, even though in
>> certain contexts in actual text an accent can exist. (I presume that
>> this is because there is no ambiguity as to where the accent would  
>> sit)
>>
>
> Here my question is: what is the intended utility of this column?  
> I'm thinking of facts such as that oblique-case forms of third- 
> declension monosylabic nouns are accented on the ultima while  
> nominatives and accusatives are accented on the root-syllable.

There are a number of reasons I've long wanted this column.

When testing my morphological parser I'm interested in the mapping  
between the lemma and this new column, not lemma and the text (which  
has additional non-morphological changes conditions by the context).

The production of an analytical lexicon from my database requires  
this new column (if you look at the Mounce or Friberg analyticals  
their non-lemma entries and normalized along the lines of my new  
column).

Having the new column for comparison with the text allows new kinds  
of queries like:
  - when does a movable-nu get dropped, or when does it stay even  
when followed by a consonant?
  - where does elision occur, or not (even though it might be expected)?
  - where are all the places where an enclitic leads to a word having  
two accents in the text?
  - find all the proclitic+enclitic or enclitic+enclitic combinations  
and categorize by how the result is accented.


> How would you list a form like POUS/PODOS with respect to the accent?

POU/S in the lemma column and for nominative singular. PODO/S, PO/DA,  
PODW=N, PO/DAS, etc.

>
>> 2. It appears that, in general, the lemmata of enclitics are shown in
>> lexicons *with* their accents. I am inclined to do so in MorphGNT.
>>
>> 2a. In the case of the indefinite TIS, BDAG accents the lemma with a
>> grave. Others don't accent it at all. What should MorphGNT do?
>>
>
> I'm not sure I see any reason to mark them; in a sense (if I  
> understand what you're doing), a context-free form of indefinite  
> TIS is purely hypothetical, existing ONLY in a dicitonary. I can't  
> see why it should be spelled as TI\S.

This is one of my key questions. Should I follow BDAG and Smyth in  
spelling it TI\S or just spell it TIS. Either way, it is an exception  
even amongst the enclitics.

I assume that it is uncontroversial to accent the oblique cases when  
listing the paradigm


>
>> 3. It appears that, in general, the context-free surface forms of
>> enclitics are shown in grammars *with* their accent except for the
>> following:
>>
>> 3a. in the case of the indefinite TIS, the accent seems to be written
>> in the context-free surface form only when it falls on the ending but
>> not when it falls on the stem (I presume that this is because the
>> accents are different in the indefinite forms than the interrogative
>> forms)
>>
>> 3b. Accents are not written on the context-free surface forms of SOU/
>> SOI/SE when enclitic (to avoid ambiguity, I presume)
>>
>> 3c. Accents are not written on the context-free surface forms of MOU/
>> MOI/ME even though there is no ambiguity with the non-enclitic  
>> variants.
>>
>>
>> What are people's thoughts on what the treatment of accents should be
>> in the context-free surface forms and lemmata of these words in
>> MorphGNT?
>>
>
> Well, it's still not clear to me what you mean the utility of this  
> column to be. This is not intended to be a lexicon or morphological  
> compendium of any sort; why is anything more than the lemma of a  
> word needed?

Hopefully the examples and uses have made it clearer. It is, in fact,  
intended to be the beginnings of a morphological compendium which is  
why I've added this new field - to more cleanly separate the  
contextual processes like elision and accent movement from the  
inflectional process.

> I do very much appreciate that your tagging of MP forms is strictly  
> in terms of morphology and doesn't attempt to interpret whether MAI/ 
> SAI/TAI or QH- forms are semantically middle or passive and that  
> you steer clear of categorizing so-called "deponents."

Credit there lies with CCAT. They did it that way in the original  
text and I have not seen any reason to change it.

> But with regard to lemmata of verbs, one might well raise the  
> question whether it would be preferable, as Randall Buth has  
> suggested, to give the aorist infinitive as the lemma rather than  
> the present indicative active 1st sg.

I do like this suggestion very much. I will definitely include it at  
some stage, the question there is whether that will take place before  
or after another large structural change I have planned: the  
splitting of the file into two: the text and a morphological database.

James
--
James Tauber                       http://jtauber.com/
journeyman of some   http://jtauber.com/blog/





More information about the B-Greek mailing list