LINGUIST List 15.888

Mon Mar 15 2004

Sum: Vocative Case/DPs

Editor for this issue: Steve Moran <>


  1. James Myers, vocative case and DPs: summary

Message 1: vocative case and DPs: summary

Date: Sun, 14 Mar 2004 02:25:38 -0500 (EST)
From: James Myers <>
Subject: vocative case and DPs: summary

Back in (Linguist 15.667) Sun Feb 22 2004, ISSN: 1068-4875), I asked
the following:

I'm mildly curious about formal analyses of the internal structure of
vocatives, but there seems to be very little research on this.

Two specific (probably unrelated) questions. First, given its role as
a discourse element that couldn't possibly be more adjunctish, how
(and more importantly why) does the vocative get case (overtly marked
in more than one language family)? Second, if proper names and
''the'' phrases are both DPs, why can only the former be used in the
vocative (again, in more than one language)? E.g. if you want the
Thing to pass you the salt, you'd say ''Hey, Thing, pass me the
salt'', not *''Hey, the Thing, pass me the salt.''

This attracted the attention of a few people more curious than I about
vocatives, case, DPs, or for that matter, syntax. I'll leave off
their email addresses so as not to feed the spambots, but interested
humans should look them up, since they have many intelligent things to
say about these topics. Thanks go to: Leo Connolly (Memphis), Joseph
F Foster (Dept. of Anthropology, U of Cincinnati), Ernest N. McCarus
(University of Michigan), Kim Schulte (School of Modern Languages,
University of Exeter), Gary H. Toops (Modern and Classical Languages
and Literatures, Wichita State University), Masha Vassilieva
(Department of Linguistics, State University of New York at Stony
Brook), R�my Viredaz (Gen�ve).

As Leo Connolly notes, it seems to be a worthwhile topic for somebody
to follow up on, but it won't be me.

(1) Regarding my first question about vocative Case, I received many
further examples (I had just been thinking of Latin and another
language I saw mentioned in a quick Google search). (I apologize if
the fonts in any of the following quotations have become messed up by
Chinese Windows on my PC.)

* Kim Schulte wrote:

...a lot of non-obligatory elements can be marked for case
overtly. Just think of all the languages with a
locative. Generativists will probably tell you that case assignment
and overt case marking have nothing to do with one another...

* R�my Viredaz wrote:

Though I have not understood why adjunctishness should logically
prevent overt case marking, I will not challenge that and I can
partially answer as far as Indo-European languages is concern. In
Ancient Greek, in the non-neuter genders of the thematic or ''second''
declension, vocative -e had in fact originally a zero ending versus
the non-zero endings -s and -n of nominative -o-s and accusative
-o-n. The origin of the o/e alternation is no longer recoverable (the
question pertains to the internal reconstruction of earlier stages of
Proto-Indo-European). From the Greek point of view, however, -e could
be seen as a non-zero ending vs. -os and -on. In the ''first''
declension, some (feminine) nouns still had a vocative in short -a
versus the nominative in -a: or -e: and the accusative in -a:n or
-e:n. In Slavic languages, in the masculine thematic declension,
nominative Indo-European *-os has yielded zero (it is disputed whether
the process is only phonological or if there has been a morphological
innovation too) and so has accusative *-om (through phonological
changes only), whereas vocative *-e has been unaltered, which is why
you now have e.g. Russian nominative Bog 'God' but vocative Bozhe (and
a new accusative Boga). In some other types of nouns, you have
e.g. Sanskrit su:nus 'son' (nom.), su:num (acc.) and su:no: (voc.),
the latter from earlier *su:nau, with zero ending but non-zero marking
however. This, too, is possibly explainable in terms of internal
reconstructive speculations.

I seem to remember having read (was it as an alleged general truth or
as an individual opinion of the writer?) that the vocative is not a
case proper. I don't remember where it was, but it was probably in
publications having a pre-generative approach. (I confess I haven't
been schooled in Generative Grammar here and I do not take G. G. and
its competing variants too seriously for the time being. As I work
mainly on historical phonology and etymology it is not very important
to make up my mind about G. G.) If vocative is not a case
syntactically, but is only treated as a case morphologically (specific
ending), there will be no logical contradiction in Government &
Binding theory, I guess.

* Leo Connolly wrote:

A few thoughts about vocatives. First, the Proto-Indo-European
vocative was endingless, and this is more or less visible in the
daughter languages which retain(ed) it as a category. Masculine Latin
o-stems had a vocative in -e, which is simply the bare e-grade of the
thematic vowel. Ancient Greek had the same ending for its o-stems,
while masculine and feminine �-stems appear with short -a
(cf. masc. nom. -�s/-�s, feminine usually -�/�).

Second, the Roman grammarians questioned whether the vocative,
despite its distinct form, should be regarded as a ''case'', since it was clearly
not part of the clause. (It could, for example, appear with sentences
lacking a second-person argument.)

These facts support an analysis in which vocative NPs precisely do not
''receive'' case. However, the Latin and Greek -e still needs an
explanation, since no case forms of o-stems actually contain a visible
-e-, so that it will not do to claim that it is an underlying form.
(That would probably have worked for PIE, however, because e/o-ablaut
was rampant there.)

There's another interesting tidbit from IE. It turns out that just as
o-stem (thematic) nouns appear with -e in the vocative, thematic verbs
have the bare thematic vowel -e in the imperative singular (Lat. mitte
'send!', Gk. l�ge 'say!' etc.) The formal identity has often been
noticed in the past, but there's no obvious explanation for it.

* Joseph F Foster wrote:

Vocative is certainly a conversation element -- I know of no culture
in which people do not address each other and no language in which
that cannot be done. But while calling to each other and labelling
the ''person'' whose attention we want is a universal, a formal
vocative is not. I use that term deliberately because I would like
for the time being to beg a little the question of what exactly it is
that a language has (must have) for us to say it has a vocative. Case
forms is of course one.

What in effect happens with a vocative is that a language has
grammaticalized a universal characteristic or property of engaging in
a conversation. In English we have ''Hey, Say, Oh, ...'' and a few
other such ''interjections'' specifically used as attention getters or
conversation starters, and there are some intonation patterns that
change a little for person's names or common nouns used vocatively,
but these sorts of things are probably universal and not what we
usually mean by a vocative. They may be a part of the ''grammar'' of
conversation, but they are not a part of the explicit sentential

I think formal vocatives may have a variety of sources,
diachronically. One of the things Ive begun looking at is
topicalization markers. For instance, the topicalization
postpositional enclitic wa in Japanese sometimes has a vocative-like
use, and what is very interesting here is that Americans who are
studying Japanese and have had a little bit of conversational practice
often use wa as a vocative in places where the Japanese do not. And
of course English doesnt have a vocative!! And topics are sort of
''adjunctish'', especially where they are in a different place from
what would normally be the subject slot.

I am pretty sure that an attention marker (Hey, Oh....) is the source
of the vocative in the Celtic languages. The marker or ''vocative
particle, is A, and it lenites, i.e. causes the ''soft mutation'' of
the initial consonant of the following word. So in Welsh

 y plant 'the children'. But 'come here, Children' is
 Dewch yma, Blant! and 'Children, come here!'' is
 (A) Blant, dewch yma!

The vocative particle is often ommitted with the result that a new
''case'' has formed and the lenition of the first consonant is then
the grammatical form of the vocative. Modern Welsh has lost all
traces of case in the nouns and most in the pronouns but it once had
them and Modern Scots Gaelic still has a vocative case and the
vocative particle, as well as a nominative, genitive, and dative case.

 Alain, 'Allen' but A Alein ' Hey Allan.
 Seaumas 'James' but A Hamish 'Hey James'.

So undoubtedly there was a true case for vocative in ProtoCeltic.

There seems to me a possibility that vocative case is an oblique form
triggered by an attention word/particle that might come to be thought
of as an oblique case triggerer like prepositions of ten are. There
is a parallel in some German dialects where wie 'like, as' has come to
trigger dative case 'wie einem Verhaltniswort'' or 'wie einer
Praposition', although in the Standard German wie doesnt trigger
dative or accusative case like many prepositions do. So in the
wie-dative dialect, they've remorphemecized and recategorized it as a
preposition. I suspect that is part of the origin of vocatives in
Celtic but dont know for certain yet. Against this proposal as a
general ''explanation'' is the fact that vocatives often have forms
unique to that case. But cf my promised attachement on Rumanian. [I
never received this -- JM]
So in sum, I think vocatives are ''overtly marked'' in conversation in
some way in nearly every langauge but it only becomes part of the
sentential grammar in some. But it is sattered around. Enda, for
instance, in New Guinea, has a vocative.


(2) Regarding my second question about determiners and vocatives, I
noted that English requires ''the'' with some proper names in ordinary
sentential context (I mentioned ''the Thing'', but perhaps more
familiar sorts of examples would be ''the Invisible Man'' or stage
names like ''the Amazing Randi'') but forbids them in the vocative.

* Kim Schulte observed:

[Determiner-less vocatives are] exactly what (normally) does happen in
Romanian, but only with masculine nouns. (So much for a separate
syntactic level...)

The masculine vocative inflection is '-e'. So the proper name 'Ioan'
becomes 'Ioane' in the vocative. But if you're addressing a boy
(baiat), a donkey (magar) or an ugly person (urit), you'll use the
vocative with the (postposed) definite article '-ul': baiatule,
magarule, uritule. ('Baiatul', 'magarul', 'uritul' mean 'the boy',
'the donkey', 'the ugly one', respectively.)

There is some fluidity in this system, but a nice example is the
minimal pair 'doamne' vs. 'domnule', both of which are regularly
formed vocative forms of ''domn'' (mister, master, lord). When you
speak to a 'mister', you use 'domnule', but when you speak to God,
'domn' is used as a proper name, so you address God as 'doamne'.

No such distinction for feminine nouns, though.

* Gary H. Toops wrote:

Just to clarify your second point: it is not uncommon to find the
definite article occurring with nouns used appellatively, whether or
not in a marked vocative form (''case''). In Canadian French it is
not uncommon to find ''le p�re'' and ''la m�re'' as forms of address
for one's father and mother, resp. French, both European and
Canadian, uses the definite article generally with appellative nouns:
''Salut, la gang!'' ('Hi, gang!'), ''Bonjour, les amis!'' ('Hello,

For more examples of this sort in French, German, Bulgarian, and a few
other languages, see Svetomir Ivanchev, ''Edna neopisana upotreba na
chlenuvata forma (k�m v�prosa za formata na obr�shtenieto v b�lgarski
ezik)'' in _Sbornik v chest na akademik Aleksand�r Teodorov-Balan po
sluchaj devetdeset i petata mu godishnina_ (Sofia: B�lgarska akademija
na naukite, 1955), pp. 271-278.

* R�my Viredaz wrote:

In French you say e.g. ''Bonjour les enfants'' (literally Hello the
children) and I even heard ''le chien'' (the dog) as a vocative
(though only in fictional contexts), when someones talk to a dog whose
name they don't know, whereas ''chien'' as vocative without the
article would be used as an insult to a man (at least in translations
of texts from other cultures), implying the information ''you are a

* Masha Vassilieva wrote:

I am working on associative plural constructions which look something
like Peter-men in Chinese and Peter-ek in Hungarian and refer to
'Peter and one(s) with' him'. Apparently, vocatives in Chinese are
also formed with '-men', which is why I'd be very interested in
knowing anything at all about vocatives and definiteness (associatives
are also always definite).

* Joseph F Foster wrote:

Well actually, one can in some languages say ''Hey, the man, pass the
salt., -- one can in naval English and one generally must in Rumanian,
hence the examples I have promised. Im looking at the time and I may
not get them to you today -- it may be a day or two, but I will get
them there. But as to the DP, I assume you mean Determiner Phrase.
That particular notation and notion is of course somewhat tied to
particular theories of grammar and how case gets formally
assigned. But I have two suggestions about languages in which one says
''Hey boys, pass the salt. and not Hey, the boys.''.
First, proper nouns are often not formally marked for
definiteness. But sometimes they are. There was/is an informal German
in which one would refer to a person, even one present, as ''der
Johann''. In Turkish a definite direct object must have a definite
d.o. case suffix, even when a proper noun. But often the propriety of
a noun is treated as redundantly definite and not morphologically or
syntactically marked.

Second, I suggest that the very act of calling to somebody has the
conversational, i.e. the pragmatic effect, of definitizing the
noun. So bachgen y"a boy' in Welsh, y bachgen y"he boy'', but
''Fachgen! '' �oy!, Hey Boy, and that is pretty definite.


(3) Regarding both questions, Ernest N. McCarus wrote:

I can give you the case of Standard Arabic, a literary language where
nouns are inflected for case (nominative, genitive, accusative) and
marked for determination (a prefixed definite article /al-/ 'the' or a
suffixed indeterminate marker /?n/ 'a').

Standard Arabic vocative nouns are nominative definite, as in /yaa
sayyid-u/ 'Sir!' (/yaa/ is the vocative particle 'O'.) Nominative is
the naming case, and is used for subjects, predicates in equational
(verbless) sentences, and for vocatives. If the person addressed is
specific, as is usually the case, the noun takes neither definite
article, which would be redundant, nor indeterminate suffix, as in the
illustration above. If it is indefinite, the classic example of which
is a blind man calling out for someone to help him, the noun is
accusative with indeterminate marker, as in /yaa rajul-a-n!/ 'O,
somebody!' ) (/rajul/ y"man")

Thus: The Arabic vocative nominal gets nominative case to show, as it
were, that an addressee is being named; if it is definite it does not
co-occur with a second definite marker but if it is indefinite it does
receive an indefinite marker and is put in the accusative, the
modifying case.


Finally, I also Googled up a couple references that seem relevant. I
haven't looked at them, though.

Moro, Andrea (2003). Notes on Vocative Case: A Case Study in Clause
Structure. In Quer, Josep, Jan Schroten, Mauro Scorretti, Petra
Sleeman, and Els Verheugd, (eds.) Romance Languages and Linguistic
Theory 2001: Selected Papers from 'Going Romance', Amsterdam, 6-8
December 2001. (pp. 247-261). John Benjamins.

Longobardi, G. (1994). Reference and Proper Names: A Theory of
N-Movement in Syntax and Logical Form. Linguistic Inquiry 25: 609-665.
[Apparently, this mentions that bare NPs may appear in the vocative in
many languages]
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue