LINGUIST List 7.765

Sun May 26 1996

Sum: Modal survey

Editor for this issue: T. Daniel Seely <dseelyemunix.emich.edu>


Directory

  1. "Alan R. King", Sum: Modal survey

Message 1: Sum: Modal survey

Date: Sat, 25 May 1996 20:39:30 BST
From: "Alan R. King" <mccayjet.es>
Subject: Sum: Modal survey
In April (issue 7.393) I posted a query on LINGUIST about root modals which
began as follows:...

The system of (root...) modality found in English, as manifested in
the modal auxiliaries 'can', 'may', 'must', 'will' etc., appears to be
unusual among the world's languages.

(1) Can anyone please show me that the assumption just stated is false
by telling me of other languages with comparable modal systems?

(2) I would also be interested in hearing about languages at the
opposite extreme, which simply do not encode root modality as
understood in English; do such exist? and if so, how do they manage?

 I got a very scanty response indeed. In April I gave it another
try. Changing strategies, I simply requested, on LINGUIST and a couple of
other language fora, translations of three sentences into as many languages
as possible:

 1) I can go to Tokyo.
 2) I have to go to Tokyo.
 3) I want to go to Tokyo.

 This new approach paid off handsomely, and within a week I received
responses covering over fifty languages. Such an excellent result having
allowed me - nay, forced me - to start up a Root Modals Data Base, I felt
that there was justification for appealing a second time round in order to
bring in yet more languages and make the whole enterprise worthwhile. In
consequence I now have data for nearly a hundred languages. Now that the
response has stabilized (I am still in correspondence with most of the
respondents - itself an exciting experience!), I am under an obligation -
which I have been urged to fulfil by several correspondents - to summarise
the results, to the extent that that is possible.

== THE INFORMANTS ==

Genuine and heartfelt thanks, first of all, to all those who answered. To
date the list of informants and consultants is as follows:

Adele Goldberg, Adrian T. Wymann, Aime Avolonto, Akeimo Birima,
Aleksander Murzaku, Alik Gilmullin, AlysseRaol.com, Ana Suffredini, Ana
Von Klopp, Anatole Lyovin, Andrew Smith, Attila Louzada, B. Koole,
Barbara Need, Bauke Koole, Bill Merrifield, Bob Thiel, Cem Bozsahin,
Christoph Eyrich, Claire Moyse, Colin Harrison, Corinne Boyle, Dana
Cohen, Dara Connolly, David Tuggy, David Wharton, David Wilmsen, Don
Davis, Dorit Ravid, Eddy Harjono, Edit Doron, Elena Bertoncini, Eleni
Manis, Elin Haf Gruffudd Jones, Elisabeth Seitz, Elke Hentschel, Emily
Ioli'i Hawkins, Ernst F. Kotze, Eulalia De Bobes I Soler, Francesca
Guardiola, Galia Hatav, Galina Briskina, Gary Holton, Geoff Smith, Ginny
Brennan, Gudmundur Erlingsson, Guy Deutscher, Haj / John Robert Ross,
Hartmut Haberland, Hilde Hasselgard, Hong Liang, Howard Gregory, Jack
Wiedrick, Jackie (Jacqueline) Murgida, Jakob Dempsey, James Kirchner,
Jan Nuyts, Jim Long, Joel Bradshaw, John E Koontz, John Lynch, John
Phillips, Judit J. Toth, Justin Smith, Karl-Michael Schneider, Keith
Goeringer, Kenneth Allen Hyde, Kevin Shapiro, Kiriaki Massoura, Kristin
Mikkelson, Leybl Botwinik, Linda Casas, Lisa Mclendon, Lynne Hewitt, M J
Hardman, M. P. Fernandez-viader, Machtelt Bolkestein, Maciej Karpinski,
Maddalena Toscano, Maik Gibson, Maite Puigdevall Serralvo, Marcia Gruss
Levinsohn, Marcia Haag, Maria, Maria Paz Buenaventura Naylor, Marja-
Liisa Helasvuo, Mark Donohue, Meiti Opie, Merce Prat Sala, Michael
Jonathan Mathew Barrie, Michel Buijs, Mikael Parkvall, Milea Angela
Froes, Miriam Meyerhoff, Nancy Stenson, Natalia Pylypiuk, Nick /
Nicholas Kibre, Nick Enfield, Noel E Rude, Norvin Richards, Olaf Husby,
Ori Pomerantz, Paul Baltes, Philippe L. Valiquette, Piripi Walker,
Rebecca Larche Moreton, Ricardo Maldonado, Robert Jones, Ronald Cosper,
Roy Iutzi-Mitchell, S Winter, Sami Laitala, Sandra Goldstein, Shamila
Naidoo, Sherman Wilcox, Silvia, Spike Gildea, Stavros Macrakis, Steve
Matthews, Steve Nicolle, Tanmoy Bhattacharya, Than-than Win, Tomas
Breathnach, Tonjes Veenstra, Tony Losongco, Twila Tardif, Uiameipulotu
Vi, Uimaitua Poloai, Ursula Doleschal, Vicky Bowman, Vincent Decaen,
Waruno Mahdi, Xulio Sousa

== OFFER OF FURTHER DATA ==

In the case of a mushrooming project of this kind, limitations of space and
time make it quite impossible to reflect the whole body of data obtained in
one summary, and given that the data continues to grow whatever I can say
now will in any case soon be out of date. Yet given the generosity with
which others have volunteered the data, I feel my own rights over the
information are limited and I have a duty to share it. So within the limits
of my own resources, I am prepared to pass on specific language data to
linguists requesting it. In making requests, please state clearly your
name, email address, and the language(s) you are interested in. There may
be a delay before the data is ready for sending, but I'll send it!

NOTE: Informants/consultants will in any case at some stage be receiving a
copy of their own languages' data with a request for them to check it for
accuracy, so are advised not to request it now to avoid useless repetition.

== THE LANGUAGES ==

These are the languages entered in my Root Modals Data Base at the time of
writing; the quality and quantity of information obviously varies from one
language to another:

AFRIKAANS; ALBANIAN; AMERICAN SIGN=ASL; ARABIC (Egyptian=Cairene;
Standard Literary; Tunisian); AYMARA; BASQUE; BENGALI=BANGLA; BISLAMA;
BURMESE; CANTONESE; CATALAN; CATALAN SIGN=LSC; CHAMORRO; CHINANTEC
(Palantla); CHINESE=MANDARIN; CHOCTAW; CORNISH; CREE (Plains); CZECH;
DANISH; DUTCH; EAST FUTUNAN; ENGLISH; FINNISH; FONGBE; FRENCH; GAELIC
(SCOTTISH); GALICIAN; GERMAN; GREEK (Classical, Modern); GUJARATI;
HAUSA; HAWAIIAN; HEBREW (Biblical, Modern); HINDI; HUNGARIAN; ICELANDIC;
INDONESIAN; INUPIAQ (Alaskan North Slope); IRISH; ITALIAN; JAPANESE;
JAQARU; KIRIBATI; KOREAN; LAO; LATIN; MANX; MAORI; NAHUATL (Copalillo;
Orizaba; Tetelcingo); NEPALI; NEZ PERCE; NORWEGIAN; NUMBAMI; OMAHA-
PONCA; PAIWAN; PERSIAN=FARSI; POLISH; PORTUGUESE (Brazilian); ROMANIAN;
RUSSIAN; SAMOAN; SARAMACCAN; SERBO-CROATIAN; SLOVAK; SLOVENE; SPANISH;
SWAHILI; SWEDISH; TAGALOG; TAMIL; TATAR; THAI; TIBETAN (Classical, Semi-
literary); TOBELO; TOK PISIN; TONGAN; TRUKESE; TUKANG BESI; TURKISH;
UKRAINIAN; VIETNAMESE; WELSH; YIDDISH; ZULU

Offers of more data for any languages not listed here will be received
gratefully (initially in the form of translations for the three sentences
given; I will follow this up with occasional further questions).

== WHAT I WANTED TO KNOW ==

Evidently, my research objective is not limited to the specific questions
asked in the queries I have posted. At the same time, the larger project to
which these queries are contributing lies outside the scope of the present
summary. But since I have been asked by several people, I shall say a few
words about the general aim.

 The specific subject of the present inquiry is one kind of modality
(a term that is notoriously vague, underdefined, and used in contradictory
or ambiguous ways). The type of modality I wish to concentrate on is that
which constitutes - according to present research - the prototypical
semantic domain of those items like _can_ and _must_ that make up the
English grammatical category variously referred to as "modals", "modal
auxiliaries" or "modal verbs". Some scholars have called this category
"root modality", some "deontic modality", and some "non-epistemic modality".
None of these terms is fully appropriate.

 The formal definining characteristics of these English modals,
notably the morphosyntactic features referred to by the acronym NICE (see my
original posting for details), are only meaningful in a highly
language-specific context and not generalisable beyond English. Not only
can we wonder whether these morphosyntactic features have any counterpart in
other languages, but more radically, whether the category of root modality
itself is cross-linguistically meaningful as a grammatical concept.
Wandering from English only as far as French, we find considerable
difficulty in maintaining this category on purely formal grounds. Hence my
first question: outside the Germanic languages (which do have quite clearcut
root-modal paradigms), how typical or atypical is the existence of the
grammatical category of root modality? This is an essential first question
prior to any broader investigation of root modals across languages, and one
subject to empirical verification.

 Shying away from conditioning the inquiry to a specific formal
theory, my approach is to take a step back from the standpoint of formal
grammar and adopt as my baseline the concept of root modality not as a
grammatical but as a notional category, and ask how different languages
express such notions; and then to examine those linguistic expressions, each
in the context of the grammatical system to which it belongs, looking for
signs of integration (or lack thereof) into a formal paradigm. In this
examination I have tried to consider the questions in the following checklist:

 (a) Are the notions of root modality systematically expressed at all?

 (b) If so, which are the most typical or least marked expressions
(where there are alternatives)?

 (c) Do these expressions display grammatical symmetry
(parallellism), asymmetry, or partial symmetry?

 (d) How much formal specialization do the one or more constructions
identified in (c) show within the grammatical system, i.e. does the language
have a formally justified specific modal construction (like English) or not
(like French, to a large extent)?

 (e) To the extent that there is asymmetry between modal expressions,
what kind of asymmetry is it, and do any similar patterns of asymmetry
repeat themselves persistently across languages?

 (f) In the case of languages giving a negative answer to (a) - if
indeed such languages exist - what kinds of communicative strategy replace
the systematic expression of root modality?

I make the initial assumption (justified elsewhere) that the different
semantic notions of *root* (deontic, non-epistemic) modality may be
conveniently grouped for practical purposes (only!) into three types, which
I refer to as POS[sibility] ("can"), NEC[essity] ("must") and VOL[ition]
("want to"). The first two are distinguishable from, albeit related to, the
notions of *epistemic* Possibility and Necessity. All these notions can be
further subdivided according to the various nuances found in one or another
language, but which it is sometimes convenient to overlook in these initial
stages of the survey at least.

== WHAT I HAVE FOUND ==

On the basis of the hundred languages examined I can state that:

 (1) An apparent minority of languages show signs that they *lack*
ordinary means for the systematic expression of the R[oot modal] concepts; I
propose to call these languages "amodal". Tentative examples: BIBLICAL
HEBREW, OMAHA-PONCA. There are also cases of "partial amodality": *some* of
the three R notions lack systematic expression, but some have it. Tentative
examples: NUMBAMI, TRUKESE. Notes: (a) this corresponds to the second
hypothesis mentioned in my first LINGUIST query; (b) this subject requires
further study, and pertinent information or ideas (for or against) would be
particularly welcome.

 (2) At the other extreme, only a small minority of languages follow
the Germanic pattern of recognising a paradigm of R notions in formal
grammatical terms to the extent of showing not only symmetry but
specialization of the R construction. The best candidate I have found
outside Germanic is CHINESE (CANTONESE and MANDARIN), and even here the
evidence is not as decisive as in Germanic.

 (3) A larger group of languages, but still only a minority, show
symmetry between the three notions without (or without very clearcut) formal
specialization. Examples include ITALIAN, POLISH, BURMESE, VIETNAMESE,
INDONESIAN and TOK PISIN.

 (4A) A much larger group of languages show partial [a]symmetry, with
two of the three basic R notions formally parallel but the third divergent.
For example: CATALAN, ENGLISH, SCOTTISH GAELIC, GREEK (Classical and
Modern), BASQUE, FINNISH, HUNGARIAN, ARABIC, FONGBE, SWAHILI, BENGALI,
HINDI, PERSIAN, LAO, TIBETAN, TOBELO, CHAMORRO, KIRIBATI, SAMOAN.

 (4B) And a fair number are fully asymmetrical, i.e. no two of the
three R notions have parallel realizations. E.g. ICELANDIC, SLOVENE, WELSH,
TURKISH, HAUSA, ZULU, JAPANESE, THAI, PAIWAN, TUKANG BESI, MAORI, TONGAN,
JAQARU, NAHUATL, NEZ PERCE.

 (5) At this point we can consider perhaps one of the most
interesting bits of the story, at least from a typological viewpoint, on
which I shall continue working: the "morphology of asymmetry", so to speak,
among languages of the two groups mentioned in point 4. All the evidence I
have collected so far seems to suggest that, of various possible features of
root modal asymmetry, some are widespread and recur again and again across
different language families and geographical areas, while others, logically
equally possible, are rare.

== ONE EXAMPLE ==

 Faced with the vastness of data, I shall limit myself to merely one
illustrative example, and one for which I didn't need informants: English
herself. Taking the following data as a basis:

POS: I can go to Tokyo. [1]
NEC: (a) I must go to Tokyo. [2]
 (b) I have to go to Tokyo. [3]
VOL: I want to go to Tokyo. [4]

we find a rather unusual kind of split symmetry here involving two
grammatical constructions distributed as follows:

 POS NEC VOL
CONSTRUCTION A [1] [2] ---
CONSTRUCTION B --- [3] [4]

Of these two constructions, A is specialized (viz. the famous NICE
properties). Construction B is not specialized: it is common to a large
number of non-modal verb/clause-complement-taking verbs which govern the
"to"-infinitive. Thus, even English fails to make the grade as a language
with a straightforward, grammaticalized and symmetrical R paradigm, of which
a much better example is e.g. DANISH.

 Although my data does not include diachronic information about
English, I believe that an earlier stage of the language would have shown,
like other present-day Germanic languages, a more symmetrical specialized
system:

POS: I can go
NEC: I must go
VOL: I will go

If this is correct, then modern English has itself diverged from the ideal
"Germanic" model with respect to its constructions [3] and [4]. (Note:
Synchronically I would not treat _will_ in present-day English as root modal
on semantic grounds, despite what has often been said. That it remains
structurally analogous to the grammaticalized root modals and thus part of
the formal "modal system" on other than semantic grounds is not being
questioned.)

 My evidence shows the evolution of an alternative, more periphrastic
NEC pattern as in [3] (cf. FRENCH _il faut_ versus _je dois_), and even the
tendency for this to partially or totally usurp the position of a more
grammaticalized competitor (cf. SPANISH, PORTUGUESE and CATALAN for the
process, and GREEK for the result), to be probably the single most
widespread phenomenon of loss of (and consequent lack of) symmetry
observable, as manifested in many resulting synchronically non-symmetrical
systems. On the other hand, the process that has led to the replacement of
"will" by "want to" is relatively rare. End of illustration.

 All comments will be extremely welcome!


Alan R. King | EMAIL: mccayjet.es
Indamendi 13, 7C | [or if all else fails] 70244.1674compuserve.com
20800 Zarautz | FAX: +34-43-130396
Gipuzkoa
Euskal Herria / Basque Country (Spain)
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue