LINGUIST List 4.212

Tue 23 Mar 1993

Sum: Adjectives as Heads

Editor for this issue: <>


  1. David Gil, Summary: Adjectives

Message 1: Summary: Adjectives

Date: Tue, 23 Mar 93 13:10:10 SSSummary: Adjectives
From: David Gil <ELLGILDNUSVM.bitnet>
Subject: Summary: Adjectives



Is anybody familiar with a language, other than English, with a
construction of the following form?

"I want the RED ONE"

I am interested in the extent to which different languages permit
adjectives to stand by themselves as complete NPs. While some
languages (eg. French, Hebrew) allow bare adjectives to occur in
NP positions, others (eg. Malay, Mandarin) make use of a
nominalizing particle or affix (eg. "yang" and "de" respectively).
English is the only language I know of that requires a dummy noun
"one". So I would be curious to find out how unique English is in
this respect.

(Also, I would be interested if anybody is familiar with any other
kinds of morphosyntactic strategies to enable adjectives to occur
in NP positions.)



The responses were as heterogeneous as are the participants of the
linguist list. Some constructions resembling the English "red
one" construction were cited in Danish, Norwegian, Gaelic, and
Breton (suggesting a possible NW European sprachbund), and,
elsewhere, Sinhalese and Ndyuka (an English-based creole of
Suriname). In other instances, constructions were offered
presenting analytical problems, in that it is not clear (at least
to me) whether they should be grouped together with the English
"red one" construction; among these are the Spanish indefinite
article (historicaly derived from the numeral "one"), the
Dravidian nominalizing suffixes (possessing many pronominal
features), and--from three respondents--the Japanese "no"
construction (which may perhaps be more closely akin to the
Malay/Mandarin nominalizing particles). Finally, a few
respondents pointed to some problems regarding the "reed one"
constrution in English.

In what follows (at some length), I have provided a minimally
edited summary of the responses, listed according to language, in
rough geographical/genetic order.


ENGLISH (Larry G. Hutchinson, David Powers, Anonymous)

English allows adjectives as NP heads fairly easily, it seems to

The carnivorous are ferocious.

She bought the blue scarves, and I the red.

Bantu languages and many typologically similar African languages
freely allow adjectives to take nominal prefixes. In Temne (West
Atlantic), for example, you can have Obai ubana "The big chief"
and Obana "The big".


Even in English, I don't think it is so clear - and it is not
clear that one is a noun here. It is clearly anaphorical, and
differs from the version where one is a noun, and we are playing
with cards with black and red numbers written on them, for

Red can stand alone : Black or Red?

Do you want the black or the red (pen)? Red. The red. The red

Anyway, I am interested in what responses you get. I would
imagine that such a usage would only occur in an uninflected
language. In an inflected language, the marking distinguishes
function - and a language like German has different systems of
inflexion for adjectives in three different contexts
(one of which is this one) and remain distinguishable from real
nouns (in at least some cases).


As only a Ph.D. student, I hesitate to contribute (and please
don't mention my name in any summary), but it seems to me that
the nature of English "one" is more problematic than your
characterization of "dummy noun" would imply. Greenbaum & Quirk
call it a "pro-form" and "nominal expression". I note the following
asymmetries between "one" and common nouns:

 I want one | * red one | * hat | * red hat
 I want * the one | the red one | the hat | the red hat
 I want * a one | a red one | a hat | a red hat

I point this out because, as a student of Japanese, I am not sure
how much difference there is between "one" and the Japanese
morpheme "no"in this context: "akai no" = "red one"
Note also the similarity in relative formation:
 [ Kim ga motteiru no ] wo mita = I saw the [ one Kim has ]


DUTCH (Jan-Wouter Zwart)

the red one: Dutch `de rode'
the red car: -- `de rode auto'

non-inflected form = `rood' (`o' and `oo' sound the same,
spelling is dependent on syllabification)

`rood' appears in the neuter indefinites:

a red book: `een rood boek'
a red one (meaning book): `een rood' or `een rode'.


GERMAN (William Edmondson)

In reply to your query. I am learning German at the moment -
slowly, and not very effectively. However, the adjective system
there is fascinating.

It is at least plausible that the analysis available for isolated
adjectives as NPs (possible in German) conforms to your
requirements. There is no need for a separate WORD, but the
adjective must carry case/gender markings not otherwise required.


DANISH (Ole Ravnholt)

Danish may be an example:

In definite nominals the head is NEVER substituted but can be

 den r|de
 the red
 "the red one"

In indefinite nominals, however, there is a (slightly
nonstandard??, but commonly used by children, at least) possibility
for substitution by EN ("one"), so that you have both

 en r|d
 a red
 "a red one"

 en r|d en
 a red one
 "a red one"


NORWEGIAN (Arnfinn M. Vonen)

Concerning your query on As as NPs:

In both standards of Norwegian (Bokmaal and Nynorsk),
referential count NPs containing an adjective and lacking a
contextually recoverable noun have an optional "dummy" that agrees
in gender with the understood noun (ex. from Bokmaal; the numeral
"one", as in English, has the same forms as the "dummy"):

SG.MASC en hvit (en) 'a white one'
SG.FEM ei hvit (ei) 'a white one'
SG.NEUT et hvitt (et) 'a white one'
PL noen hvite (noen) 'some white ones'

SG.MASC den hvite 'the white one'
SG.FEM den hvite 'the white one'
SG.NEUT det hvite 'the white one'
PL de hvite 'the white ones'

The semantic difference between, e.g., "en hvit" and
"en hvit en" is elusive, perhaps it is mainly stylistic;
the dummy seems to be preferred in the singular in colloquial
speech. Norwegians with intuitions in standard-like dialects may be
able to help you there.

More "exotic", perhaps, are the following paradigms
from my native (non-standard) dialect of the Trondheim area.
There, indefinite singulars seem to require the "dummy";
in the plural, however, zero is OK
(beware: informal transcriptions!):

SG.MASC en kvit *(en) 'a white one'
SG.FEM ei kvit *(ei) 'a white one'
SG.NEUT et kvitt *(et) 'a white one'
PL non kvite (*non) 'some white ones'

The definite paradigm, on the other hand, shows another strategy:

SG.MASC kvitn (*en) 'the white one'
SG.FEM kvita (*ei) 'the white one'
SG.NEUT kvite (*et) 'the white one'
PL kvitan (*non) 'the white ones'

No pre- or postposed elements are required,
and the A takes a special set of suffixes. These look like the
definite suffixes of the most productive declensional classes of nouns,
but cannot be identified with the suffix of the understood noun, since
some nouns take other suffixes, whereas the adjective suffix is
invariable. E.g., "kvit-e bil-an" 'the white cars' and "kvit-e mus-en"
'the white mice' both correspond to the N-less NP "kvit-an"
'the white ones', whereas there is no *"kvit-en".

If there are several adjectives in the phrase only the last one gets
the special "nominal" suffix: "ny-e kvit-an" 'the new white ones'.
The A with the "nominal" suffix may still take adverbial modifiers:
"aller kvitest-an" 'the very whitest ones'.

If you find this somehow interesting, please feel free to
ask for details. I also look forward to knowing what you find
out about the topic in general.


GAELIC (Caoimhin P. ODonnaile)

 Scottish Gaelic:
 Tha mi ag iarraidh an fhear dearg (masculine nouns)
 Be I a' seeking the masculine thing red

 Tha mi ag iarraidh an te dearg (feminine nouns)
 Be I a' seeking the feminine thing red

 Irish Gaelic:
 Ta me ag iarraidh an cheann dearg (inanimate nouns)
 Be I a' seeking the thing red

I have glossed
 fear masculine thing
 te as feminine thing respectively
 ceann thing

but really the situation is this: "fear" means "man" and "ceann"
means "head" (in both languages), but in Scottish Gaelic "fear" and
"te" are used as sorts of pronouns, according to whether the object is
grammatically masculine or feminine and regardless of whether it
is animate or inanimate. In Irish Gaelic, however, the split is
according to whether the object is animate or inanimate, "ceann"
being used for the former and "te" for the latter, regardless of sex.

The word "iarraidh" means "seeking, requesting, actively wanting".
A more passive kind of desire or need is expressed differently.
However, I don't think this is what you are interested in, and the
"dummy nouns" remain the same.

I haven't bothered to indicate acute accents in the spelling, and
the fine details of the grammar (use of genitive case) may be wrong
or subject to argument. I can check the details of the grammar if you
need this. There's a list WELSH-LIRLEARN.UCD.IE where you (or I,
if you like) could inquire about the P-Celtic languages. I only know
about Gaelic.


BRETON (Greg Stump)

Breton is parallel to English in this regard. There is a dummy noun
_hini_ (plural _re_) that is essentially like English _one(s)_:
_an hini ruz_ `the red one', _ar re ruz_ `the red ones'.


SPANISH (Bill Robboy, Jeff Runner)

In Spanish, `I have a/one red dress' is

 Tengo un vestido rojo.
 I.have a dress red

where _un_ is the normal masc. sg. form of the article/numeral
`a/one' in prenominal position.

`I have one', with identity-of-sense anaphora with a masc. sg.
antecedent, is

 Tengo uno.
 I.have one

where yu have _uno_, the form also used for masc. sg. `one' as a
predicate (`we are one'), a pronoun with generic (right term?)
reference (`one does what one can'), or a mathematical object.

`I have a red one', again with identity-of-sense anaphora with a
masc. sg. antecedent, is

 Tengo uno rojo.

The following are bad:

 *Tengo uno vestido rojo.

 *Tengo un.

 *Tengo un rojo.

The facts are similar with at least a couple of other determiners
(and pronouns?) whose masc. sg. forms are shorter in prenominal
contexts than other contexts.

 Tienes algu'n vestido rojo?

 `Do you have any red dress?'

 Tienes alguno?
 `Do you have any (masc. sg.)?'

 Tienes alguno rojo?
 `Do you have any red one (masc. sg.)?'

 *Tienes alguno vestido rojo?

 *Tienes algu'n?

 *Tienes algu'n rojo?

 No tengo ningu'n vestido rojo.
 `I don't have any red dress.'

 No tengo ninguno.
 `I don't have any (masc. sg.).'

 No tengo ninguno rojo.
 `I don't have any red one (masc. sg.).'

 *No tengo ninguno vestido rojo.

 *No tengo ningu'n.

 *No tengo ningu'n rojo.

(u' here represents u with an acute accent over it. Orthographically
there should also be an upside-down question mark at the beginning
of the questions.)

I should say that I'm not a native Spanish-speaker and haven't
actually sat down and checked all this with one, but I know the
language quite well and these facts are pretty elementary ones that
should be easily verifiable. On the face of it, they would seem to be
susceptible to more than one analysis, but at least it's clear that you
can't simply plug in an adjective in place of a nominal -- at least not
without some morphosyntactic adjustments.


in response to your message on linguist, i think that spanish has
the kind of construction you're interested in. as far as i know the
following sentence is correct (i'm not a native speaker):

quiero el rojo
want-1sg the-masc red-masc
'i want the red (one).'

the adjective can be "nominalized" by using the definite
determiner. i don't know much more about it than that, though.


GREEK, ANCIENT (David Brandt)

In ancient Greek, the distinction between noun and adjective is
very blurry, since the substantive use of the adjective is extremely
common. At any place in a sentence, all an adjective has to do to
function as a noun is simply stand by itself. (In such an instance, it's
usually translated with a bland noun, such as "man" or "thing.")


BULGARIAN (Grace Fielder)

If I understand your question correctly, you would like to know
that Bulgarian allows the adjective to stand alone, although in the
meaning of The Red (One), it will have the definite article:
Iskam krasnija (where -ija is a definite article). The South Slavic
Languages have this definite:indefinite distinction, which is
manifested differently in the other (North) Slavic group.


RUSSIAN -- see Mandarin Chinese


HUNGARIAN (Istvan Kenesei)

Of course Hungarian is one of the lanuages you're looking for, cf.

Mari a fekete kalap-ot vette meg, Anna pedig a piros-at.
M. the black hat-ACC bought PERF A. however the read-ACC
'Mary bought the black hat, and Anna the red one.'


SINHALESE (John C. Paolillo)

Sinhala, an Indo-Aryan language spoken in Sri Lanka, also
uses the numeral "one" as a nominalizer for adjectives. Tamil,
a Dravidian language, uses a form _itu_, etymologically a pronoun.
Imbabura Quechua (see Peter Cole's Monograph)
has a somewhat different strategy which may indicate
an intermediate status for Adj -- bare adjectives can bear
nominal markings (e.g. accusative case) in nominal positions,
but in adjectival positions they do not.


TAMIL -- see Sinhalese


DRAVIDIAN (Kodama Nozomi)

In reference to your QUERY: ADJECTIVES dated 16/03/93
Dravidian languages make use of a series of nominalizing affixes
to have a complete NP without its head noun. The entire process is
often referred to as pronominalization since these affixes are
almost homophonous with the distant demonstrative pronouns.
Following examples are from Telugu:
naa "my", naa-di "mine" f&, adi "that, she" f&
erra "red", erra-vi "red ones", avi "those, they"
nii-laaNTi (oka abbaayi) "(a boy) like you",
nii-laaNTi-vaaDu "someone like you", vaaDu "that man, he"
puli-ni camp-ina-vaaLLu 'tiger-ACC kill-PP-m&'
"those who (will have) killed the tiger", vaaLLu "they" m&
erra-vaaTi(-loo) "(among) red ones", vaaTi "they" OBL
mana-laaNti-vaaLL-am "people like you and me (incl. us)"m&
 *vaaLL-am cf. (manam) hinduvul-am "we Hindus"
As seen from the examples, these affixes are not mere nominalizers
but markers of inflexional categories i.e. gender, number, case
and, for NOM, person-number, which all require to be specified for
any NP.
The English dummy functions similarly in this respect. The
difference in morphological status may be ascribed to the fact that
English allows coordinated modifiers in NPs while Telugu does not.
I sometimes suspect that the so-called head noun of an NP may be
modifiers of the real head, Spec.



Hi. Your query was of interest to me because I've just completed
writing a section of my dissertation which deals extensively with
some of the data that you mentioned, namely the Mandarin 'de' in
post-coplula position. My work actually deals with this material
from a quite different viewpoint, that is, I'm primarily interested in
the morpheme 'de' itself, which I claim is a morpheme that marks (in
GB terms, 'heads') a modifier, and does not function as a nominalizer.
Nevertheless, as an example of the phenomenon which your
concerned with, those data are quite appropriate. You might be
interested to know that there is a significant (or at least detectable)
difference in meaning between the two copula constructions of
Mandarin, the one with and the one without the copula. In the latter,
where 'de' is not present, the adjective is understood as 'absolutive',
in the sense that itsdenotation holds of the subject in absolute terms:
Mandarin 'that insect big' means that the insect is large in an
absolute sense. In the other copula construction, where there is an
overt copula and 'de', themeaning is relative. Mandarin 'that insect
shi big-de' means that it is a big insect, even though it might be quite
small in absolute terms. This might be explained if we hold that
there is a null nominal head in the shi-de construction which picks
up its denotation from the context, or the subject, or in some other,
broadly construed anaphoric way. The null nominal head theory also
gives us a natural way of predicting the presence of 'de', and not
some other 'nominalizing morpheme', under the analysis of DE which
I propose in my dissertation: it is 'de' precisely because all modifiers
of nouns, even null ones, are marked with 'de'. Interestingly, a
similar distinction in interpretation occurs when PPs occur in the two
copula constructions.

I got the idea to look at this distinction in meaning in Chinese
when reading about the so-called long and short form adjectives in
Russian. Ican't remember any of the specific citations right off the
top of myhead, but Leonard Babby, Muffy Seigal (sp?) and John
Bailyn (a collegue of mine here at Cornell) all discuss the difference
in meaning mentioned above, and basically they all conclude that the
long form adjectives of Russian, when in predicate position, modify a
null noun. So Russian has some data for you.


JAPANESE (Peter Austin, David J. Silva, Robert Westmoreland) --
see English

Japanese is a language of the type you seek. It has an empty
nominal 'no' that is used with adjectives to head an NP, eg.

akai no ga suki desuo
red one subj likable copula
"I like the red one"

Some adjective-like elements in Japanese are nominal-adjectives.
These take 'na' before 'no', as in:

suki na no wa akai no desu
likeable na one topic red one copula
'The one I like is the red one"

This 'no' can also serve to head NPs with nominal modifiers.
Usually when a nominal modifies a head (eg. in a posessive
construction) the genitive particle'no' appears - you don't get two 'no'

suzuki san no ga suki desu
Suzuki Mr/s one subj likeable copula
"I like Mr/s Suzuki's one"

Notice that this 'no' empty nominal can also head relative clauses
in Japnese.


I don't know a whole lot about Japanese, but as I recall, there
are constructions such as "akai no" 'red one' and "takai no"
'high/tall one' and "furui no" 'old one', where "no" is some sort of
dummy N that supports the Adjective. An analogous situation holds
for Korean (whichI have a much better handle on) in phrases like
"ppalgan gEs" 'red one' and 'cohUn gEs' "good one" -- here [gEs]
(where "E" = schwa or open 'o') is a morpheme meaning "thing". It
also is used as a dummy noun, both in these adjectival constructs and
in sentential constructs:

wurinun hakkyoe mos ganun gesi cohci anhta.
we school-to not going THING good not
"Our not going to school is not good" or better yet
"The fact that we can't go to school is not good"

These are suggestions from me--of course, I'd check 'em out w/
native speakers.


Japanese has a parallel constuction (although the parallel may not
be immediately obvious in a translation of your example sentence
because the way one expresses "I want X" is quite different!)

 Akai no ga hosii.

akai (adj) = 'red'
no (noun?) = 'one'
ga ("particle") = subject/nom case marker
hosii (adj) = 'want'

Here are some more sentences:

 Mina akai no o katta.
 'Everybody bought red ones.'

mina (pronoun?) = 'everybody'
o (pcl) = direct obj/acc case marker
katta (verb, past) = 'bought'

 Hurui no wa yasui.
 'The old one is cheaper.'

hurui (adj) = 'old'
wa (pcl) = topic marker
yasui (adj) = 'cheap'


KOSRAEAN (Martin Haspelmath)

here+s another example of a de/yang-type element. It comes from
Kosraean (Lee, Kee-dong. 1975. Kusaiean reference grammar. UP of
Hawaii.), p. 379:

the word is ma:

Pahpah el ahsack ik na luhlahp soko a ninac el ahsack ma na
srihksrihk soko.
father caught fish big mother caught "one" small
"Father caught a big fish but mother caught a very small one."

Interestingly, ma is also used to introduce relative clauses, like

Oak soko ma nga orwaclah ah tihli.
canoe that I made sank
"The canoe that I made sank."


NDYUKA (George Huttar)

Ndyuka, an "English-based" creole of Suriname, uses _wan_ 'one'
afteradjective in a structure parallel to the English one:

Mi wani a lebi wan. 'I want the red one.'
1s want the-sg red one

_wan_ in this environment is unstressed (and carries low tone), as
it does prenominally when functioning as indef. sg. art. It is
stressed (and has high tone) when it is the numeral 'one'.

Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue