Clitic Properties
Chris
I have just posted a new set of clitic properties for open review.
The idea is that we will leave the properties posted on this group for
a month or so, in order for people to comment on them (in this
thread).
The authors are:
Andrea Cattaneo and Oana Savescu Ciucivara
(with comments from Chris Collins, Stephanie Harves,
Richard Kayne and Paco Ordonez)
Chris
hilda
I started adding some comments in the document file itself, but will
take them out.
and paste them somewhere separately, before posting them.
comments will be forthcoming..
--hilda
hilda
Here are some of my comments (preceded by ** or comment, boldface or
color have disappeared).
Happy holidays!
--hilda
**It would be good to have HELP notes that can expand in the
definitions in case language experts encounter specific problems. Some
of my comments fall under this heading**
**Maybe it would be good (for the future?) to have a general table/or
tables of the properties, so that it is clear all cases are included;
I assume they are; I am not checking for missing cases…**
**Suppose a language does not have clitics at all, or has one object
clitic. Would it be good if the language expert would not have to fill
out N/A for each of the properties below (i.e. should we automatise a
bit?) or would we miss cases. Would it still be interesting to have
examples for each of the cases below for languages without clitics (I
guess yes)?**
** For increased readability, how about adding here and everywhere
below, between parentheses: Cl.1sg ((to)me) Cl.2sg.DO (you). (I
personally have always intensely disliked these glosses, and would be
strongly in favor of some changes along these lines).
**In some languages (Maasai, inverse languages, the action is between
subjects and objects. (1/2)/ 3 on 1/2 DO). (Often these are fused
forms). It would be good to expand properties in the future to include
these**
**The properties do not distinguish between 3 inanimate and 3 animate/
person. I think they should. **
1. Clitic definitions:
Indirect object and direct object clitics are variants of strong
object pronouns. Such clitics are unstressed, need a phonological host
and are typically adjacent to the verb.
**change: weaken ‘typically' to often? (cf Dutch clitics are not in
the vicinity of a V; it is not even clear what they are phrasing with
phonologically ).
Their syntactic distribution is different from that of strong
pronouns and regular DPs (cf. Kayne 1975).
add illustration? Jean voit Marie ..Jean la voit
** [In quite a few lgs 3rd clitics are zero; or inanimate 3rd is zero,
animate is a clitic. any ideas how to hunt for zeros? ]
** Comment: In some traditions, pronominal clitics are called object
agreement markers. My guess is we want ‘clitic’ to cover these.
I propose illustrating for Bantu [and adding a help note] about
terminology? (few Bantu languages can have multiple clitics)
3. Parameters regarding word order:
Clitic Verbfinite
The property Clitic Verbfinite has the value "yes" when there is at
least one clitic that can immediately precede the finite verb.
Verbfinite can be either a finite lexical verb or a finite auxiliary.
**comment1: Finiteness is often not an obvious notion.. (maybe in the
help modue: verb finite: the verb used in a simple statement, and
illustrate?)
**comment2. In some languages, clitics always precede the lexical
verb, regardless of auxiliary like elements. (Bantu) [SA T ] [cl V…].
I guess we want to have these come out as yes.
**Comment 3. what about second position clitic languages? What do we
expect here? (Neutral contexts will be subject (DP) initial, and hence
these will be yes?)
**Possibility: splitting the property in 2 properties. i.e. separate
finite lexical verb from finite auxiliary? [future: clitic climbing/
restructuring?)
Cl.1SG. IO Cl.2SG.DO in Finite Contexts
** how about adding here and below, between parentheses
Cl.1sg ((to)me) Cl.sg.DO (you) (that makes it so much easier to read)
**How do we get languages with fused morphemes? we definitely want a
yes answer.
Cl.2SG.IO Cl.3SG.DO in Finite Contexts
Example
Romanian is ‘yes’ for Cl.2SG.IO Cl.3SG.DO in Finite Contexts:
Mama ţi
l - a
dat ieri.
Mother-the cl.2sg.io cl.
3sg.m.do has
given yesterday
‘My mother gave it to you yesterday’.
**Need to control for animacy of DO;
**split property in two? one for inanimate and one for person?
Cl.3SG.IO Cl.3SG.DO in Finite Contexts
The property Cl.3SG-IO Cl.3SG-DO in Finite Contexts has the value
"yes" when there is at least one instance where such a sequence is
licit in a finite context either preverbally, or postverbally. We
restrict our attention to neutral contexts, in which there is only one
lexical verb, and to productive word order patterns.
Definition of Clitic
General Note on Word Order Properties: A "yes" value for Cl.3SG-IO Cl.
3SG-DO in Finite Contexts does not exclude a "yes" value for any
other clitic sequence.
NA (Not Applicable): If a language does not have overt clitics, this
property has the value NA.
Example
Romanian is ‘yes’ for Cl.3SG.IO Cl.3SG.DO in Finite Contexts:
Maria i l
- a prezentat.
Mary cl.3sg.io cl.3sg.m.do
has presented
‘Mary introduced him/her to her/him’.
** animate
Slovenian is ‘yes’ for Cl.3SG.IO Cl.3SG.DO in Finite Contexts:
Janez mu ga
je še dal.
Janez cl.3sg.io cl.3.sg.do aux.3.sg still give.prt
‘Janez still gave it to him’
**inanimate
Comment: There are languages which, in the sequence cl.3sg.io cl.
3.sg.do, exhibit less transparent clitic forms (see Perlmutter 1971
for Spanish and Jones 1993 for Sardinian).
---------------------