Table Of ContentINVESTIGATING LINGUISTIC ACCEPTABILITY
JANUA LINGUARUM
STUDIA MEMORIAE
NICOLAI VAN WIJK DEDICATA
edenda curat
C. H. VAN SCHOONEVELD
INDIANA UNIVERSITY
SERIES MINOR
NR. LIV
1966
MOUTON & CO.
• THE HAGUE •
LONDON PARIS
INVESTIGATING
LINGUISTIC
ACCEPTABILITY
by
RANDOLPH QUIRK and JAN SVARTVIK
LONDON GÖTEBORG
1966
MOUTON & CO.
LONDON • THE HAGUE • PARIS
© Copyright 1966 by Mouton & Co., Publishers, The Hague, The Netherlands,
No part of this book may be translated or reproduced in any form, by print,
photoprint, microfilm, or any other means, without written permission from
the publishers.
Printed in the Netherlands
ACKNOWLEDGMENTS
This study reports work conducted within the Survey of English
Usage and as such it is a product of collaboration between all the
members of the Survey staff. To our colleagues Henry T. Carvell
(on whose help with the statistical work we have heavily relied),
Derek Davy, Jo Evetts, Norman Fairclough, Judith A. Godfrey,
Sidney Greenbaum, and Joan Mulholland, therefore, we wish to
express our special gratitude for their co-operation from the plan-
ning of the first experiment to the completion of this report. In
addition, we have profited from valuable help given by others at
various stages of the work: the skilled advice of John K. Hall in
analysing and presenting the data; the help of Eric Quirk in check-
ing and correcting; the specialised criticism of Peter C. Wason
from a psychological point of view. In the early stages of planning
the experiment, we received important help and advice from Noam
Chomsky, during his visit to London in November 1964. The
debt we owe to predecessors and fellow-workers in the field of
linguistic acceptability is, we trust, made sufficiently plain by the
references throughout, but we should like to acknowledge addi-
tional help in private communications from David G. Hays (to
whom we owe Fig. 10), Paul Ziff (who made valuable comments on
a preliminary version of this study), D. L. Bolinger, and Barbara
M. H. Strang.
University College London R. Q.-J. S.
September 1965
CONTENTS
Acknowledgments 5
1. Introduction 9
2. Test Design 18
3. Informant Accuracy and Homogeneity 26
4. The Operation Test 35
5. The Judgment Test and its Relation to the Operation Test 49
6. Linguistic Grouping of Test Sentences 67
7. The Selection Test 84
8. Conclusion 97
Tables 105
References 118
1
INTRODUCTION
'Grammaticalness' involves two prime but interrelated difficulties -
establishing what it is and determining native reaction in respect
of it, the interrelation entering through the obvious fact that the
second is dependent on the first, the linguist's categorial problem.
On the first of these, one writer has summed up the position in a
light-hearted but fundamentally serious comment: 'I have been
unable to find any agreement among modern linguists as to what
constitutes a grammatical sentence. At one extreme there are those
who call every utterance a sentence, that is any string of words
ever mouthed by poet or peasant. At the other extreme there are
those who would declare cannibalism ungrammatical on the
grounds that man does not belong to the class of food-nouns'
(Lambek, 1961, 167). The second difficulty has been epitomised
with similar epigrammatical irony: 'Psycholinguistic material is
notoriously hard to manage. Naive native speakers of Standard
Average European are almost extinct' (Hays, 1964, 525).
A great deal of interest has been focussed on the problem in
recent years, and it is beyond question that this interest has been
aroused through the startling clarity with which 'the fundamental
aim' of a grammar has been formulated by generative linguists:
that it should account for 'all and only the grammatical sentences
of a language' (cf. Chomsky, 1957, 13, 21; Bach, 1964, 5). What
is startling here is that a tacit assumption traditional in all gram-
mar-writing has been made explicit, and consequently we are
confronted with the full implications of this assumption. The
most important of these for our present purpose is not that there
are the entities 'fully grammatical' and 'fully ungrammatical' sen-
tences but that there is the ability to draw a line between the two.
The confrontation makes it clear that grammatical rules which
10 INTRODUCTION
have hitherto been preoccupied with preventing the kind of devi-
ance of which native speakers are capable
(1) Him and her don't want no cake.
or of which foreign speakers are capable
(2) I am here since two years.
must also explicitly prevent the production of sentences that neither
natives nor foreigners would probably ever be in danger of forming,
these sentences themselves being of at least two further types:
(3) Little a boy the ran street up.
(4) Colourless green ideas sleep furiously.
Already at this point the categorial difficulty becomes acute. That
is, while no one would deny that a description of English must
explicitly exclude (3) and (4) and that the native speaker has no
difficulty in placing both (3) and (4) comfortably beyond the pale
of acceptable sentences, plenty are ready to deny that 'grammar'
is involved in excluding both of them. Nor would this seem to
be a trivial matter of the traditional way in which the word 'gram-
mar' has been used either in linguistics or in popular (for example,
educational) practice - even if the latter could reasonably or safely
be dismissed as trivial. The ability to see (4) as in some important
ways perfectly grammatical (and differing from ordinary sentences
only in being concerned with nothing that anyone would want to
say) is matched by the reluctance to see its deviance as attributable
to ungrammatical usage and by a preference instead to regard it
as merely 'ontological nonsense' (cf. Bolinger, 1961, 371). Atten-
tion has been drawn to the fact that a 'distinguished philosopher
maintains that "He had a green thought" is grammatical but
nonsensical' (Ziff, 1964, 210), and it seems relevant to point out
that such widespread reactions as those of the distinguished phi-
losopher coincide significantly with the fact that there is little
evidence that the philological tradition of any language brings
deviations of the type found in sentence (4) within a grammatical
description. (See further, pp. 65f.)