summary: demonstrative or pronoun

David Gil gil at EVA.MPG.DE
Fri Aug 14 03:04:09 UTC 2009

Dear all,

Here is a summary of the responses to the "demonstrative or pronoun" 
query I posted last week. But first a reminder of the query:


Consider the following very similar contexts;

Context A:
John and Bill are friends. John calls Bill on a landphone; it's a bad 
line, Bill doesn't know who is speaking; John tries to identify himself 
(using a predicate nominal construction)...

Context B:
John and Bill are friends. John sends Bill a text message from a new 
number that Bill is unfamiliar with; John identifies himself (using a 
predicate nominal construction)...

My question:

In languages that you are familiar with, in the above contexts, is the 
subject of the predicate nominal construction a demonstrative or a 1st 
pronoun pronoun?

In English, the subject is a demonstrative; the pronoun is infelicitous 
in the given context:

This is John
#I am John

But in Indonesian, the subject is most commonly a pronoun, though a 
demonstrative is also possible:

Ini John [less common]
Aku John

I am curious to know what happens in other languages. (I have a hunch 
that the availability of the "pronominal subject" option in Indonesian 
is correlated with the questionable status of pronouns as a discrete 
grammatical category in Indonesian, but this hunch is easily testable 
with a bit of cross-linguistic data.)

Note: I don't expect to find differences between the two contexts; I 
provided both just in order to make the situation more natural to as 
many respondents as possible.


Below, languages are classified as Type 1 if they use a demonstrative 
(with a variant Type 1a for the use instead of a locative expression); 
and as Type 2 if they use either a 1st person singular pronoun or else a 
copula inflected for the 1st person singular. Note that Type 2 languages 
may fall into two subtypes depending on whether the construction is more 
like "(I) am John" or more like "(It's) me, John"; although these are 
quite distinct constructions, it was not always possible to tell apart 
from the responses, and so the two were collapsed. (Indeed, for 
Indonesian I would not be able to distinguish the two myself.) In 
addition, some languages are shown as being of mixed type, allowing more 
than one construction type, or in the "other" category, allowing neither 
of the construction types under consideration.

Type 1: Languages that say "DEM is John"
English, Dutch, Norwegian, Polish, Hebrew, Sakha

Subtype 1a: Languages that say "LOC is John"
French, German (Swiss), German, Czech, Macedonian, Arabic (Tunisian), 
Arabic (Standard), Japanese

Type 1 / Subtype 1a: Languages that say both "DEM is John" and "LOC is John"
Finnish, Sami (Northern)

Type 2: Languages that say "(I) am John" / (It's) me, John
Basque, Spanish, Italian, Greek, Hungarian, Arabic (Yemenite) [other 
preferred], Arabic (Lebanese/Syrian/Palestinian) [other preferred], 
Georgian, Armenian, Azeri, Persian, Hindi, Mongolian (Khalkha), Korean, 
Cantonese, Menya (Angan, PNG)

Type 1/2: Langauges that say both "DEM is John" and "(I) am John" / 
(It's) me, John
Russian [Type 1 preferred], Mandarin [Type 2 preferred], Indonesian 
[Type 2 preferred], Jamaican Creole

Tamil, Kalapalo (southern Carib, central Brazil), Papiamentu


My hunch that the Indonesian 1st-person-pronoun usage was somehow 
related to the problematical nature of pronouns in Indonesian and other 
Southeast Asian languages was comprehensively refuted within minutes of 
posting the query -- see the large number of typologically diverse 
languages of Type 2 above. But for what it's worth, that was the only 
clear result to emerge from the query.

As a miniature typological survey, the resulting sample suffers from an 
overwhelming Eurasian bias, and probably also from inconsistency: 
without further investigation, it is not always clear that an apparent 
pairwise difference is between languages rather than observers. Still, 
within (greater) Eurasia, there would seem to be a rough areal pattern, 
with Type 2 languages extending from PNG across southern Eurasia to the 
Basque Country, and Type 1 languages to their north. If this pattern is 
real (something that is not at all certain), then it must be either the 
result of extremely recent (post Alexander Graham Bell) and hence rapid 
diffusion, or else the product of ancient spread, via contact, of some 
deeper grammatical property that gives rise to the distinction, whatever 
that may be.

Some respondents mentioned additional constructions that would be 
appropriate in the given contexts, while others proposed additional 
discourse contexts that would be associated with different 
constructions. For example, in English, in response to a question such 
as "Who are you?", "I am John" is clearly appropriate. Another issue 
that came up was the reliability of native-speaker judgments as opposed 
to naturalistic data. (I should point out that what prompted this query 
was a piece of naturalistic data: receiving a text message in Indonesian 
from an old friend who I had been out of touch with, which began with 
"Aku [proper noun]", or 'I am [proper noun]')

Thanks to my fellow typologists who responded to the query: Peter 
Arkadiev, Jon Aske, Dik Bakker, Ellen Basso, Winfried Boeder, Bernard 
Comrie, Anaid Donabedian, Viktor Elšik, Pål Eriksen, Joseph Farquharson, 
Jocelyne Fernandez-Vest, Victor Friedman, Gideon Goldenberg, Dolgor 
Guntsetseg, Claude Hagége, Alice Harris, Hakyung Jung, Siva Kalyan, 
Olesya Khanina, Silvia Kouwenberg, Bingfu Lu, Silvia Luraghi, Stephen 
Matthews, Annie Montaut, Edith Moravcsik, Samia Naim, Miren Lourdes 
Oñederra Olaizola, Brigitte Paakendorf, Johannes Reese, Anna Siewierska, 
Don Stilo, Hannu Tommola, Peter Trudgill, Nigel Vincent, Bernhard 
Waelchli, Carl Whitehead.

David Gil

Department of Linguistics
Max Planck Institute for Evolutionary Anthropology
Deutscher Platz 6, D-04103 Leipzig, Germany

Telephone: 49-341-3550321 Fax: 49-341-3550119
Email: gil at

More information about the Lingtyp mailing list