Professional Documents
Culture Documents
≥
Brought to you by | The University of Texas at Dallas (UTDALLAS)
Authenticated
Download Date | 4/24/17 6:41 PM
Studies on Language Acquisition
18
Editor
Peter Jordens
Mouton de Gruyter
Berlin · New York
edited by
Michel Achard
Susanne Niemeier
Mouton de Gruyter
Berlin · New York
앪
앝 Printed on acid-free paper which falls within the guidelines
of the ANSI to ensure permanence and durability.
ISBN 3 11 017357 3
Bibliographic information published by Die Deutsche Bibliothek
Die Deutsche Bibliothek lists this publication in the Deutsche Nationalbibliografie;
detailed bibliographic data is available in the Internet at ⬍http://dnb.ddb.de⬎.
쑔 Copyright 2004 by Walter de Gruyter GmbH & Co. KG, D-10785 Berlin.
All rights reserved, including those of translation into foreign languages. No part of this book
may be reproduced in any form or by any means, electronic or mechanical, including photocopy,
recording, or any information storage and retrieval system, without permission in writing from
the publisher.
Cover design: Sigurd Wendland, Berlin.
Printed in Germany.
Index 281
group of people
operating together
(clandestinely)
'----r---.,-----' mmmmmj ''On, I
..............-
3. A Usage-based Model
The proposal would thus be that the child initially learns individual, item-
based linguistic constructions (e.g., verb island constructions), and ifthere
are patterns to be discerned among these different item-based constructions
in adult usage, she could then make abstractions and create inheritance hier-
archies of constructions. In this view of language and its acquisition, there-
fore, there is continuity not of structures - adults control a more diverse and
abstract set of constructions than do children - but there is continuity of
process in the sense that the processes of learning and abstraction are the
same wherever and whenever they are applicable ... This general approach is
usage-based in the sense that all linguistic knowledge - however abstract it
may ultimately become - derives in the first instance from the comprehension
and production of specific utterances on specific occasions of use.
the polysemous "over" where the multiple senses associated with the prepo-
sition are organized in a connected network of related meanings. Tyler and
Evans further argue that their approach enables teachers to rely on the well-
known meanings of the central senses to gradually move on to the most
peripheral ones, thereby strengthening their semantic motivation.
The second language research program inspired by cognitive linguistics
is still in its infancy, and a lot of the results presented here are preliminary.
However, it is our hope that the different chapters of this volume will help
establish the cognitive linguistics model as a valuable framework for the
investigation of second language learning and teaching phenomena, and
provide the methodology to further extend the research.
References
Teresa Cadierno
1. Introduction
The question of how adult second language learners come to express spatial
relations in a second language is a rather neglected area within second
language acquisition research (but see Becker and Carrol1997 for an excep-
tion).l Within first language acquisition research, on the other hand, extensive
research has been conducted on how children perceive and express spatial
relationships in their native language (see Bloom et al. 1996 for a recent
overview). While earlier research focused on the influence of the child's
nonlinguistic knowledge on the development of spatial language (e.g.,
Piaget and Inhelder 1956; Levine and Carey 1982), recent cross-linguistic
approaches under the general framework of functional-cognitive linguistics
(e.g., Berman and Slobin 1994; Bowerman 1994; Bowerman 1996a & 6b)
have examined the ways in which children from different languages initially
use spatial language.
This cross-linguistic research, which has compared same-age children
learning languages with different spatial categories (e.g., Choi and Bowerman
1991; Bowerman 1994) has shown that the way in which children initially
classify space for language is the outcome of a complex interplay between
the child's nonlinguistic spatial knowledge and the input of the language
that he/she is learning. The conclusion reached by Bowerman (1996a) is
that the influence of the input language is stronger than it was believed. In a
similar way, cross-linguistic research conducted by Berman and Slobin
(1994) within a cognitive typological framework (Talmy 1985) has shown
that children learning typologically different languages differ with respect
to how much and what kind of information they provide when referring to
particular spatial situations involving motion.
As expressed by Bowerman (1996b), the hypothesis that language can
influence children's construction of semantic categories has been rather
unpopular given its associations with the controversial Whorfian hypothesis,
according to which the way in which human beings view reality is shaped
by the particular language they speak. In order to account for the above-
mentioned findings in cross-linguistic research and at the same time decouple
such issues from the loaded Whorfian hypothesis, Slobin (l996a) has coined
the expression "thinking for speaking" to refer to the influences of language
on the ongoing process of interpreting and expressing verbal messages.
According to Slobin (1996a), the results of the cross-linguistic research in
first language acquisition conducted by Slobin and his colleagues (Belman
and Slobin 1994) support the hypothesis that when acquiring his/her first
language, a child learns specific ways of thinking for speaking.
The purpose of the present investigation is to extend this line of research
conducted within L1 acquisition to the field ofL2 acquisition. More specifi-
cally, the study examines the implications of Slobin's thinking for speaking
hypothesis for L2 learners within a cognitive typological approach, and thus
investigates how L2 learners come to express spatial relationships involving
motion in a language - Spanish - that is typologically different from their
native language - Danish.
2. Conceptualization in Language:
the Thinking for Speaking Hypothesis
while the last two are associated components: (a) Motion: the presence per
se of motion; (b) Figure: the moving entity; (c) Ground: the object with
respect to which the Figure moves, which can potentially include the source,
the medium, and the goal of movement; (d) Path: the course followed by the
Figure with respect to the Ground; (e) Manner: the manner in which the
motion takes place; and (t) Cause: the cause of its occurrence.
Talmy (1985: 61) offers the following two examples to illustrate the
components:
In both sentences the pencil functions as the Figure and the table as the
Ground which in this example expresses source of movement, while the
particle off expresses the Path. In addition, the verbs rolled and blew express
Motion while in addition the verb rolled expresses Manner of motion while
blew expresses Cause of motion.
According to Talmy (1985), there are crucial cross-linguistic differences
in the ways in which different languages of the world express motion
events. This author has identified two main typological patterns in which
the universal components of motion events are expressed by different
means in different languages: (a) satellite-framed languages (S-languages),
such as Chinese and all branches of the Indo-European family except
Romance Languages, where the main verb conflates Motion + Manner /
Cause, and Path is expressed separately by means of a satellite, i.e., verb
particles (e.g., down, out, up, etc) and prefixes (e.g., mis- as in misfire).
Some examples in English are: The rock slid / rolled / bounced down the
hill (Motion + Manner); The napkin blew off the table (Motion + Cause);
and (b) verb-framed languages (V-languages), such as Romance, Semitic,
and Polynesian languages, where the main verb conflates Motion + Path,
and Manner and Cause are expressed separately by means of an adverbial or
a gerundive type of constituent. In some cases, the expression of this con-
stituent may result in an awkward style, so that it is often omitted. An exam-
pie in Spanish would be: La botella entro / salio de la cueva (flotando) (The
bottle entered / exited the cave (floating)).
In their choice oflanguages, Berman and Slobin (1994) together with his
associates used Talmy's typology selecting S-languages (i.e., English and
Gemlan) and V-languages (i.e., Spanish, Hebrew, and Turkish). The results
of the study showed consistent typological differences in the way motion
events were narrated. More specifically, children leaming typologically dif-
ferent languages differed strikingly by age 3 in the way they talked about
motion. These differences, which resulted in distinct rhetorical styles, can
be summarized as followed:
locative description 'below was the water' allows one to infer that the
trajectory went from some elevated place to the pond. As shown by the
two examples, the two language types suggest a differential allocation of
attention between description of movement (i.e., details of trajectories)
and description oflocation (i.e., static descriptions).
5. Speakers of S-languages provided more manner information than speakers
of V-languages. This is mainly due to the fact that S-languages have a
larger lexicon of manner of motion verbs, which allows their speakers to
make finer manner of motion distinctions than speakers of V-languages.
In other words, each language "trains" its native speakers to pay different
kinds of attention to particular details of events when talking about them.
This is possible given that "the language or languages that we learn in
childhood are not neutral coding systems of an objective reality; rather,
each one is a subjective orientation to the world of human experience, and
this orientation affects the ways in which we think while we are speaking"
(Slobin 1996a: 91). That is, the ways in which native speakers of different
languages express and describe motion are largely determined by the lexi-
calization patterns of their native language.
The implication of this "training" for second language acquisition has
been pointed out by Slobin himself (1996a) who hypothesizes that the train-
ing received while learning our native language may turn out to be rather
resistant to restructuring in adult second language acquisition. The present
study is presented as an attempt to examine the implications of Slobin's
thinking for speaking hypothesis for adult second language learners whose
The general research question that guided the present study was the follow-
ing: How do L2 learners come to express motion events in an L2 that is
typologically different from their Ll? More specifically, how do adult lan-
guage learners whose Lt can be characterized as an S-language (Danish)
come to express motion events in a V-language (Spanish)? The general
hypothesis is that the learners' Lt-thinking for speaking patterns will be the
point of departure for the interpretation and production of L2 patterns, and
that, consequently, learning a second language will involve learning another
way ofthinking for speaking, that is, learning how the semantic components
of a motion event are mapped into L2 surface forms, and learning which
particular details of a motion event must be attended to in the input and
expressed in the L2.
In the case of the present study, the Danish learners of Spanish must
learn: (l) the characteristic L2 form-meaning mappings, which are different
from the Lt mappings, i.e., Path typically mapped in the verb, Manner typi-
cally mapped in a separate expression; and (2) the (subtle) differences
between the Ll and the L2 with respect to the particular aspects or details of
the motion event that must be focused on: (a) relatively less elaboration in
descriptions of Path of motion, i.e., less use of ground adjuncts; (b) relatively
less frequent use of event conflation; (c) relatively more narrative attention
to location-static descriptions vs. movement-trajectories; and (d) relatively
less elaboration of Manner of motion.
The present study, which focused on the semantic components of Path
and Ground, addressed the following specific research questions:
Applying the typology suggested by Talmy (1985), Danish and Spanish rep-
resent two different typological patterns. Danish is a Germanic language
charactet1zed as an S-language, where Path is typically expressed by an elab-
orate system of satellites, i.e., verb particles which indicate an opposition
between translocative and locative function (e.g., ud-ude 'out (translocative)-
outside', ind-inde 'in (translocative)-inside'; Han gik ud 'He went out' vs.
Han var ude 'He was outside'). Manner and Cause of motion are typically
encoded in the verb (e.g., kravle 'crawl', snige sig 'creep', smide 'throw').
Tn fact, Danish could be characterized as a more prototypical example of an
S-language than English in that it lacks Latinate verbs (e.g., enter, ascend,
descend), thus preventing the conflation of Motion and Path in the verb
root.
4.2. Method
4.2.1. Participants
higher proficiency level consisting of the two learners who had studied and
lived in a Spanish-speaking country, plus six learners who had studied
Spanish in Danish high schools as well as in Spanish-speaking countries for
long periods of time (between seven months and one year). A total of eight
learners fell into this group. The proficiency level of these learners would
be comparable to the advanced proficiency level described in ACTFL
guidelines on writing; and (2) a subgroup of learners with a lower profi-
ciency level, consisting of the remaining eight learners who had studied
Spanish in Danish high schools but who had either not been in a Spanish-
speaking country, or had only been there for very short periods of time
(under two months). The proficiency level of these learners would be com-
parable to the intermediate-mid proficiency level described in the ACTFL
guidelines on writing.
Finally, the native speaker group consisted of university students from
Spain who had just arrived in Denmark for one semester's study as part of
the European Socrates interchange program. At the time of data collection,
they all reported no knowledge of Danish.
The experimental procedures for the native speaker group were identical to
those of the learner group with two exceptions: (a) they only performed the
task in Spanish; and (b) they were not presented with a bilingual list of
nouns.
4.3. DataAnalysis
The written data were first divided into clauses which, following Berman &
Slobin (1994: 660), were defined as "any unit that contains a unified predi-
cate". By "unified" was meant a predicate that expressed a single situation,
and it included both finite and nonfinite verbs. In general, clauses consisted
of a single verbal element, but infinitives and participles functioning as
complements of modal or aspectual verbs were included with the matrix
verb within a single clause (e.g., querer ir 'want to go', empezar a caminar
'start walking').
Once the written texts had been divided into clauses, the clauses
expressing motion events were identified, i.e., motion clauses. These were
the clauses that formed the basis for further analyses. Following Slobin
(1996b: 206), a motion event was defined as " ... the description of the
movement of a protagonist from one place to another. ..." The criterion was
thus that the protagonist had to end up in a different place from where he
started. Thus, for example, in the sentence Om natten da drengen sover i sin
seng,jlygter freen ('In the night when the boy sleeps in his bed, escapes the
frog'), two clauses were first identified - Om natten da drengen sover i sin
seng and jlygter jreen - and only the second clause was subsequently iden-
tified as involving a motion event.
The following analyses were conducted on the data: (1) a comparison of
the length of the texts produced by the two subject groups, i.e., the Spanish
native speaker group and the L2 learner group. This analysis was conducted
to examine whether the narratives written by the two groups were compara-
ble in this respect; (2) a type/token analysis of the motion verbs found in the
Danish and Spanish data; (3) an analysis of the degree of elaboration of
path of motion. This analysis comprised: (a) an examination of the satellites
found in the Danish and the Spanish data; and (b) an examination of the
extent to which the Danish and the Spanish data included minus-ground vs.
plus-ground clauses; (4) an analysis of the extent to which the Danish and
the Spanish data included event conflation; and (5) an analysis of the rela-
tive allocation of attention to movement (trajectories) vs. location (static
descriptions) in both languages.
In the analyses the following comparisons were made: Ca) Spanish narratives
written by the Spanish native speaker group, i.e., Ll Spanish narratives; Cb)
Danish narratives written by the learner group, i.e., Ll Danish narratives;
and Cc) Spanish narratives written by the learner group, i.e., L2 Spanish nar-
ratives.
results seems revealing. In addition, it was observed that the Spanish native
group included numerous instances of dialogues and conversations between
the main characters, either in direct or indirect speech. Such instances were
rare in the narratives produced by the Danish group.
Given the significant differences in total length mentioned above, a sec-
ond analysis was conducted which examined the relative frequency of
motion clauses used, i.e., clauses encoding motion events. That is, this
analysis examined the amount of motion clauses used in relation to the total
number of clauses produced by both subject groups. A Wilcoxon signed-
rank: test performed on the three means (Ll Spanish: 30.8; Ll Danish: 34.9;
L2 Spanish: 28.5) revealed no significant differences in the performance of
the two subject groups in their Ll, i.e., the Ll Spanish narratives and the Ll
Danish narratives (p = 0.1134). Additionally, no significant difference was
found between the Ll and the L2 Spanish narratives (p = 0.4509). In other
words, a comparable relative frequency of motion clauses was obtained by
(a) the Spanish NS group and the Danish learner group performing in the
Ll; and (b) the Spanish NS group and the Danish learner group performing
in the L2.
To summarize, the narratives written by the Spanish speakers were sig-
nificantly longer than the narratives written by the Danish speakers in both
Danish and Spanish. This difference, which was attributed to the inclusion
of additional information not strictly related to the chain of motion events
depicted in the "frog story" picture book, did not, however, have an effect
on the relative frequency of motion clauses used by the two subject groups.
Thus we can be confident as to the validity of the following analyses, which
are based solely on information contained in the motion clauses.
This analysis examined the total number of tokens and types found in the
Danish and the Spanish data. The motion verbs used by the learner group in
both Danish and Spanish are presented in Appendix A and the motion verbs
used by the Spanish NS group are presented in Appendix B. The lists
include both verbs of self-motion (e.g., caminar 'to walk') and of caused-
motion (e.g., tirar 'to throw'). Furthermore, following Slobin (1996b), a
plus sign following a verb indicates that it has been used with one or more
satellites.
The learner group used a total of 63 verb types in Danish and a total of
37 verb types in Spanish, whereas the Spanish NS group used a total of 67
verb types in Spanish. Table 1 shows the mean values for the tokens and
types used by the two subject groups along with the token-type ratios.
Table 1. Tokens, types, and token-type ratios in Danish and Spanish: Mean values
A Wilcoxon signed-rank test conducted on the mean scores for the verb types
in isolation revealed the following: (a) there were no significant differences
in the amount of verb types produced by the Danish and the Spanish speak-
ers in their Lt (p = 0.1659); (b) the learner group produced significantly
more motion verb types in the Lt-Danish than in the L2-Spanish (p =
0.0012); and (c) the learner group used significantly fewer motion verb
types in Spanish than the native speakers of this language (p = 0.0001).
Another Wilcoxon signed-rank test was conducted on the mean scores
obtained for the token/type ratios. This analysis differs from the previous
one in that it examines the verb types used by the subject groups in relation
to the total number of verb tokens produced, and it thus provides a more
balanced description of the data. The results of this analysis parallel those of
the previous analysis with respect to (a) (p = 0.9099) and (b) (p = 0.0410).
With respect to (c), however, the difference in performance between the
learner group and NS group did not reach significance (p = O. 0568).
However, the low p value obtained here in combination with the significant
result from the previous analysis seems to point to a difference in the amount
of Spanish motion verb types used by the learner and the NS groups.
In sum, the results obtained in the present analysis support the first
hypothesis posited at the outset of the study, namely, that the L2 learner
group would use fewer motion verb types in Spanish than the NS group.
Second language acquisition entails a process of progressive vocabulary
learning, and it was expected that these learners, although belonging to the
intermediate and advanced stages of language acquisition, would still not be
able to use as wide a variety of motion verbs as native speakers of this lan-
guage. It was noted that, for example, learners would use only one verb
(e.g., ir 'go') whereas Spanish native speakers would use additional verbs
with similar semantic import, such as dirigirse 'go' and encaminarse 'be on
one's way'. Furthermore, the difference in the Ll-L2 performance by the
learner group was also to be expected. That is, it was expected that the
learners would have at their disposal a greater variety of verb types in their
LI-Danish than in the L2-Spanish.
What was unexpected was the result concerning the non-significant dif-
ference between the two subject groups in their L1. This result seems to
contradict previous research by Berman & Slobin (1994) and Slobin (1998)
which has shown that native speakers of S-languages produced more verb
types than native speakers of V-languages. The results of the present inves-
tigation indicate that even though the two language types may differ with
respect to the total amount of motion verb types contained in their lexicons,
the subject groups investigated here showed no difference as to their use. A
possible explanation for the difference in results between Slobin's study and
the present study may be that whereas the former used oral narratives as
elicitation method, the latter used written narratives. Stylistic differences
between the written and the oral mode of communication, which are well
documented both at the grammatical and the lexical level (e.g., Schiffrin
1987; Vigara 1992), may explain the different results of the two investiga-
tions. Thus, at the lexical level, it is to be expected that the native speakers
will use a greater variety of motion verbs in the written mode than in the
oral mode, and that consequently the differences between native speakers of
S- and V-languages found in oral language will be leveled out in the case of
written language. A detailed analysis of the Spanish native data showed that
each individual tended to use several motion verbs with similar meanings in
their narratives (e.g., andar and caminar 'walk'; tirar and arrojar 'throw';
huir and escapar 'flee'). If the narratives had been oral, the speakers might
not have produced such a wider variety of lexical units encoding motion.
This is a matter that deserves future investigation.
Satellites in the Danish and the Spanish data. Under (1 a) is the complete
list of Danish verbs + satellites contained in the Danish data:
(la)
(lb)
'The dog fell out of the window', where the adverb (a)fuera specifies the path
of motion followed by the Figure - the dog. This type of construction, which
involves boundary-crossing situations was not found in the data from the
Spanish native speakers who, instead, favored constructions such as El perro
se cayo de la ventana 'The dog fell off the window' where path of motion is
omitted. The constructions found in the learner data thus seem to indicate that
the learners are adding some extra semantic information, namely information
on the path of motion, by means of a satellite. This semantic information,
on the other hand, must be obligatorily expressed in the learners' Ll-Danish
by a satellite, as in Sd hoppede hunden 1!:.d of vinduet 'So jumped the dog out
of the window' / Hunden faldt 1!:.d af vinduet The dog fell out of the window'.
Consequently, the presence of this construction in the Spanish L2 learner
data seems to indicate a process ofL! transfer at work. More specifically, in
accordance with the hypothesis posited at the outset of this study, some
learners, i.e., those who are at the intermediate stages of language acquisi-
tion, do exhibit a higher degree of complexity and elaboration of the path of
motion in Spanish than the native speakers of this language. This degree of
elaboration is, in turn, parallel to that found in the learners' Ll-Danish.
In addition to the anomalous constructions discussed above, the L2
learner data included another type of anomalous construction which is
exemplified by sentences such as El ciervo mueve al niiio y a su perro abaiQ
en un precipicio 'The deer moves the boy and his dog down in the abyss'
and Cuando el chico intenta irse arriba de algunos arboles 'When the boy
tries to go up of some trees'. In these examples, the learners have used a
non-directional verb of movement (i.e., mover 'move', ir 'go') followed by
a satellite encoding vertical path (i.e., abajo 'down', arriba 'up'). In addi-
tion, in the first example, the learner has omitted the semantic component of
cause which is present in the native Spanish rendering of the sentence, as in
El ciervo tiro / arrojo al niiio y a su perro a un precipicio 'The deer threw
the boy and his dog in the abyss' .
There are two possible explanations for all these types of anomalous
sentences, which are in agreement with some of the views proposed in the
literature on transfer: 5 (a) the learners are inadvertently transferring an L 1
pattern to the L2, possibly as the result of shared conceptual systems under-
lying both their Ll and their L2 structures (as discussed in Slobin 1993 and
Jarvis 2000, among others); or (b) the learners are evidencing a type of
communication strategy; that is, when the learners are confronted with a
communicative situation where they lack the knowledge of the appropriate
Spanish verb of movement, they purposely use an Ll structure to fill in the
gaps in their knowledge of the L2 (as suggested by Corder 1983, among
The present analysis examined the extent to which the narratives produced
by the L2 learners and the Spanish native speakers included clauses incor-
porating a ground specification. As indicated by Slobin (1996a), descrip-
tions of movement can be equally enriched in both S- and V- languages by
means of prepositional phrases which add locative details to the verbs, and
which can express the ground of movement, i.e., source, medium, or goal:
'The boy and the dog hid themselves behind a tree trunk. '
The aim of this analysis was to examine whether the Danish learners of
Spanish tended to make use of event conflation, i.e., to investigate whether
from those described in the work of Slobin and his colleagues in that not
only two locations (the cliff-source and the water-goal) were mentioned, but
also a third one, the so called afgrund 'abyss'. This inclusion was very
interesting since its use seemed to invoke different conceptualizations of the
scene, depending on the grammatical construction in which it appeared, an
issue that will be discussed below.
Ll Danish 2 7 2 5
Ll Spanish 3 13 o o
L2 Spanish 2 11 3 o
Notes: I loc.= Reference to one location; 2 lac. = Reference to two locations;
3 loco = Reference to three locations; E. C. = Event Conflation.
(b) Reference to two locations - either the cliff, the abyss or the water - in
separate clauses:
Hjorten lober mod en afgrund / hvor den stopper hrat / og drengen faldel' ud
over og /led i vandet
The deer runs towards an abyss / where it stops suddenly / and the boy falls
out over and down in the water
These two Danish examples are interesting to compare because they show
that the same word, i.e., the afgrund 'abyss' seems to invoke two different
conceptualizations when used in different grammatical constructions. Thus, in
the first example, the word afgrund 'abyss' seems to profile6 the undefined
endpoint of the downward movement; that is, the boy is profiled as moving
on a path downwards from an implicit unstated source - the cliff - towards
an abyss which is pictured as being below that source. And in the next
clause, the endpoint of that abyss is defined as the water. In the second
example, on the other hand, the word afgrund 'abyss' seems to profile the
starting point or origin of the downward movement; here the boy is profiled
as moving on a path from the beginning of an abyss, from which he falls
down into the goal-ground, the water. In other words, in the first example,
the word afgrund focuses on the lower part or the depth of the abyss, and as
such it is conceptualized as an undefined goal, while in the second example
it focuses on the top part of the abyss, and as such it is conceptualized as a
sort of origin or beginning of the downward movement.
The same patterns in the conceptualizations are found in the Spanish data:
(c) Reference to three locations: the cliff, the abyss and the water, in separate
clauses:
/fiorten anbringer drengen i sit gevir I og gar imod en steil skrt:ent I hvor den
lader drengenfalde ned i gfgrunden I Hunden og drengen lander samtidigt i
vandet.
The deer places the boy in his antler / and goes towards an abrupt cliff /
where he lets the boy fall down in the abyss !The dog and the boy land at the
same time in the water.
Hjorten lager ham med / og smider ham ud over en afg,rund ned i en brfJnd
The deer takes him with / and throws him over an abyss down in a pond.
In this example, the second clause specifies two locations, the abyss and the
pond.
As shown in Table 3, the predominant patterns of expression for the Lt
Danish narratives were (a) the reference to two locations in separate clauses
(7 subjects); and (b) event conflation (5 subjects). The remaining subjects
chose either to make reference to the three locations or only one of the loca-
tions (the goal). In contrast, the predominant pattern of expression for the
Lt Spanish narratives was the reference to two locations in separate clauses
(13 subjects). The remaining 3 Spanish native speakers made reference to
only one location (the goal). With respect to the L2 Spanish narratives, the
predominant pattern of expression was the reference to two locations in two
separate clauses (11 learners). The remaining learners either made reference
to three locations or only one location.
A Fisher's exact test performed on these data revealed the following: (a)
a significant difference was found between the patterns followed by the two
groups in their Lt (p = 0.020). This result mirrors the one found by Berman
& Slobin (1994) concerning the different rhetorical styles followed by adult
speakers of S- vs. V- languages: whereas the former tend to favor event con-
flation (i.e., the most compact construction) and event serialization, i.e., the
expression of two or three locations in separate clauses, the latter tend not to
use event conflation, favoring instead event serialization; and (b) no signifi-
cant differences were found between the Spanish native speakers' patterns
and the Spanish L2 learners' patterns (p = 0.258). This result does not support
the hypothesis stated at the outset of the study, namely that there would be
differences in the way in which Danish learners of Spanish and native
speakers of this language would use event conflation in their nalTatives. As
shown in Table 3, the L2 learners investigated here did not employ event
conflation and consequently, did not incorporate the typological pattern
characteristic of their mother tongue into their second language.
A possible explanation for this finding may be what Kellerman (1978,
1986) has termed "psychotypology", which refers to the learners' perceived
distance between the constructions in the two languages. Kellerman demon-
strated that the L2 learners have some perceptions about a) what structures
from their native language are or are not prototypical and therefore poten-
tially transferable or non-transferable, and b) whether the native and target
languages are relatively similar or different. These perceptions influence
what they actually transfer into the L2. In the case of the present study, it
could be the case that the Danish learners perceive constructions involving
event conflation as characteristic of Danish and uncharacteristic of Spanish,
thus constraining or preventing transier. This is a matter that deserves further
investigation.
1 + +
2 + +
3 + +
4 + +
5 + +
6 + +
7 + +
8 + +
9 +
la
11
12
13
14
15
16 +
Total 9 8
L 1 patterns and the Spanish L2 patterns (p = 0.500). This result does not
support the hypothesis posited at the beginning of this investigation, namely,
that there would be differences in the relative allocation of attention to
movement and location by Danish learners of Spanish and Spanish native
speakers. It is important to note, however, that this result is due to the mixed
pattern of use followed by the Spanish native speakers; and Cc) a significant
difference was found between the learners' performance in the L 1 and in the
L2 (p = 0.0156). That is, whereas the learners followed the typical pattern of
S-languages, i.e., allocation of attention to trajectories, in their LI-Danish,
their pattern of use in the L2-Spanish was mixed.
Table 4 further indicates different patterns with respect to the learners'
performance in the Ll and the L2. Ten learners allocated differential
attention to movement and setting in the Danish and the Spanish versions of
the frog story. Nine of them (No. 1-9) provided descriptions of trajectories
in the Ll-Danish and descriptions of settings in the L2-Spanish, whereas
the other learner (No. 16) followed the reversed pattern, i.e., she provided a
setting description in the Ll-Danish and a description of the trajectory
involved in the L2-Spanish. The following example illustrates the former
pattern, which reflects the characteristic rhetorical style of the two typologi-
cally different languages involved in this study:
Ved hjrelp af sit stor gevir bar hjorten drengen hen til en skrrent. / hvorefter
den smed ham ned i afgrunden. / Hunden ... rag samme vej / Heldigt for
dem var det / at de faldt ned i en dyb sa.
With the help of his big antler carried the deer the boy to a cliff / where after
he threw him down in the abyss / The dog ... went the same way / Luckily
for them was / that they fell down in a deep pond.
In the Danish version, the segments of the path are explicitly encoded by
means of a satellite - ned - and the setting thus needs to be inferred: i.e., the
fact that the cliff is located above the lake. In the Spanish version, on the
other hand, the setting is explicitly stated. That is, it is explicitly stated that
there is a lake below at the bottom of the abyss, a description which allows
the path to be inferred.
The remaining six learners (No. 10-15) followed the same narrative style in
both languages, i.e., they specified the segments of the path, thus leaving
the setting to be inferred by the reader, as in the following example:
Men /ige for skrcellten naede ~jorten at stoppe /og bade Peter og Frede
faldt ned i afgrunden, men heldigvis - lige i en skovso.
But just before the cliff, managed the deer to stop / and both Peter and Frede
fell down in the abyss - but luckily in a forest pond.
The fact that one informant (No. 16) provided a static description in the
Danish version together with the mixed pattern of use exhibited by the
Spanish native speakers included in this study suggests that the rhetorical
styles characteristic of the two language types are a result of strong tendencies
in the languages, thus allowing for possible overlapping across typologies:
5. Conclusions
AppendixA.
Danish verbs:
anbringe 'place', baske 'flap', begive sig+ 'move', bremse+ 'stop', bringe
'bring', brere+ 'carry', drage+ 'pull', dukke+ 'emerge', falde+ 'fall', fange
'catch', fare+ 'rush', flygte 'escape', flyve+ 'fly', forf£llge 'chase', forlade
'leave', forsvinde+ 'disappear', f£llge+ 'follow', fH 'get', fa + past partici-
ple+, galopere+ 'gallop', gemme sig 'hide', gH 'walk', havne 'land', hente
'fetch', hoppe+ 'jump', jagte 'chase', kaste+ 'throw', klatre+ 'climb',
komme+ 'come', kravle+ 'crawl', lande+ 'land', liste sig+ 'creep', lregge
(sig) 'lay', lrene sig+ 'lean', l£lfte+ 'lift', l£lbe+ 'run', medbringe 'bring
along', medtage 'bring with', myldre+ 'swarm', nrerme sig 'approach', na
'reach', ramme 'fall on', rende+ 'run', ryge+ 'fly off', ryste 'shake', samle+
'assemble', skubbe+ 'push', slippe+ 'release, let fall', srnide+ 'throw', snige
sig+ 'creep', snuble+ 'stumble', stige+ 'rise', stikke+ 'put, thrust', stille sig+
'place', stoppe+ 'stop', st£ltte sig+ 'lean', stii+ 'stand', srette (sig)+ 'place',
tage+ 'take', undslippe 'escape', vandre 'stroll', vrelte+ 'knock over', vrere
pa vej 'go'.
Spanish verbs:
acercarse 'approach', acostarse go to 'bed', aparecer 'turn up', asomar
'begin to appear', aventar 'throw', caer(se) 'fall', capturar 'capture', coger
'take', correr 'run', dejar 'leave behind', dejar caer 'let drop', desaparecer
'disappear', encerrar 'shut in', escalar 'climb', escapar(se) 'escape', escon-
der(se) 'hide', hacer caer 'make fall', huir 'flee', ir(se) 'go', levantarse 'get
up', llegar 'arrive', llevar(se) 'carry', meter 'insert', mover 'move', parar
'stop', perseguir 'chase', poner 'put', resbalar 'slide', sacar 'take out', salir
'leave', saltar 'jump', seguir 'follow', subir(se) 'ascend', tirar 'throw', venir
'come', volar 'fly', volver 'return'.
AppendixB.
Spanish verbs:
abandonar 'leave', acercarse 'approach', acostarse 'go to bed', adentrarse
'go into', alejarse 'go away', andar 'walk', aparecer 'appear', apoyarse
'lean', arrastrar 'drag along', arrojar(se) 'throw', asomarse 'lean oneself
over', atravesar 'cross', caer(se) 'fall', caminar 'walk', capturar 'seize',
coger 'take', colocar 'put', correr 'run', dejar 'leave', depositar 'put', deten-
erse 'stop', derrapar 'skid', desaparecer 'disppear', dirigirse 'go', empujar
'push', encaminarse 'be on one's way', enganchar(se) 'hook up', entrar
'enter', escapar(se) 'escape', esconderse 'hide', flotar 'float', hacer caer
'make fall', huir 'flee', introducir 'insert', ir 'go', ir de vuelta 'go back', lan-
zar 'throw', levantarse 'rise', llevar(se) 'transport', llegar 'arrive', mar-
charse 'leave', merodear 'maraud', meter 'insert', moverse 'move', parar
'stop', pasar 'go by', perseguir 'pursue', poner pies en polvorosa 'flee',
echarse para atras 'back out', menear 'waggle', precipitarse 'precipitate',
recoger 'pick up', regresar 'return', refugiarse 'take refuge', retirarse 'go
away', sacar 'take out', salir 'leave', saltar 'jump', seguir 'follow', subir(se)
'ascend', tirar 'throw', transportar 'transport', trepar 'climb', venir 'come',
volar 'fly', volcar 'capsize', volver 'return'.
Appendix C.
Notes
1. This research was funded by Aarhus University Research Foundation and The
Faculty of Humanities at the University of Southern Denmark. Some of the
issues discussed in this paper were presented in a talk by Teresa Cadierno and
Anne Jensen at the AAAL conference in Vancouver (2000). I would like to
thank Werner Vach for his valuable statistical help, as well as Eva Dam Jensen
and Lucas Ruiz for comments on earlier versions of this article. Thanks also to
Sharon Millar for linguistic revisions of the paper.
2. Given that previous research on interlanguage variability (e.g., Tarone 1983) has
shown L2 learners' performance to be more target-like in written tasks than in
oral tasks - the reason being that in written tasks they can attend more closely to
the language code - written data were collected in this study. Differences in
results between previous research and the present one, which could be attributed
to differences in elicitation mode, are discussed later in the paper.
3. This sentence was also produced by a few adult native speakers of Spanish in
Sebastian and Slobin's (1994) study. Its use was accounted for by the fact that
the situation describes a kind of non-canonical change of location, which makes
the native speaker emphasize the fact that the dog had actually put his head
inside the jar.
4. Another possible explanation offered is that the two groups of presehoolers, i.e.,
the ones that used the directional adverbs and the ones that did not, simply re-
present two different types of individual styles. Further research is needed to
shed some light on this issue.
5. I would like to thank Nick Ellis for bringing this issue to my attention.
6. The term "profile" has been used by Langaeker (1987, 1998) to refer to a kind of
prominence involving the focusing of attention: "Within the full conceptualization
it evokes, an expression directs attention to some particular substructure - its
profile - as being the entity that it designates or refers to. An expression's profile
is thus its 'referent', in a psychological ... sense of that term." (Langacker 1998: 9).
7. I would like to thank an anonymous reviewer of this paper for his/ her comments
on this regard.
References
Andersen, R.
1990 Models, processes, principles and strategies: Second language acqui-
sition inside and outside of the classroom. In B. VanPatten & J. F. Lee
(Eds.), Second language acquisition - Foreign language learning.
Clevedon, Avon: Multilingual Matters.
Becker, A. & M. Carrol
1997 The acquisition of spatial relationships in a second language
Amsterdam / Philadelphia: John Benjamins.
Berman, R. A. & D. I. Slobin
1994 Development of linguistic fonns: English. In R. A. Berman & D. I.
Slobin (eds.). (1994). Relating events in narrative: A crosslinguistic
developmental study. Hillsdale, New Jersey: Lawrence ErIbaum.
Bloom, P., M. A. Peterson, L. Nadel, & M. F. Garret
1996 Language and space. Cambridge, Mass: The MIT Press.
Bowerman, M.
1994 From universal to language-specific in early grammatical develop-
ment. Philosophical Transactions of the Royal Society. London,
B34637-45.
1996a The origin of children's spatial semantic categories: Cognitive versus
linguistic determinism. In J. J. Gumperz & S. C. Levinson (eds.),
Rethinking linguistic relativity (pp. 145-176). Cambridge: Cambridge
University Press.
1996b Learning how to structure space for language: A crosslinguistic
perspective. In P. Bloom, N. A. Peterson, L. Nadel & M. F. Garrett
(eds.), Language and space (pp. 387-436). Cambridge, Mass: The
MITPress.
Cadierno, T. & K. Lund
In press. Cognitive linguistics and second language acquisition:
Motion events in a typological framework. In B. VanPatten, J.
Williams & S. Rott (eds.), Form-meaning connections in second
language acquisition. Hillsdale, New Jersey: Lawrence ErIbaum.
Choi, S. & M. Bowerman
1991 Learning to express motion events in English and Korean: The influ-
ence of language-specific lexicalization patterns. Cognition 41:
83-121.
Corder, P.
1983 A role for the mother tongue. In S. Gass & L. Selinker (eds.),
Language transfer in language learning (pp. 85-97). Rowley, MA:
Newbury House.
Ellis, R.
1994 The study ofsecond language acquisition. Oxford: Oxford University
Press.
Ibarretxe-Antuiiano, I.
2001 Motion events in Basque narratives. Linguistic LAUD Agency. Series
A. General and Theoretical Papers N. 528. Universitiit Essen.
Jarvis, S.
2000 Methodological rigor in the study of transfer: Identifying L1 influence
in the interlanguage lexicon. Language Learning 50 (2): 245-309.
Kellerrnan, E.
1978 Transfer and non-transfer: where are we now? Studies in Second
Language Acquisition 2: 37-57.
1986 An eye for an eye: crosslinguistic constraints on the development of the
L2lexicon. In E. Kellerman & M. Sharwood Smith (eds.), Cross-lin-
guistic influence in second language acquisition. Oxford: Pergamon.
Langacker, R. W.
1987 Foundations of cognitive grammar, Vol 1: Theoretical perspectives.
Stanford, California: Stanford University Press.
1998 Conceptualization, symbolization, and grammar. In M. Tomasello
(ed.), The new psychology of language: Cognitive and functional
approaches to language structure. London: Lawrence Erlbaum.
Levine, S. C. & S. Carey
1982 Up front: The acquisition of a concept and a word. Journal of Child
Language 9: 645-657.
Lucy, J. A.
1992 Grammatical categories and cognition: A case study of the linguistic
relativity hypothesis. Cambridge: Cambridge University Press.
1996 The scope of linguistic relativity: An analysis and review of empiri-
cal research. In J. J. Gumperz & S. C. Levinson (eds.), Rethinking
linguistic relativity (pp. 37-69). Cambridge: Cambridge University
Press.
Mayer,M.
1969 Frog, where are you? New York: Dial Press.
McClure, E., M. Mir, & T. Cadierno.
1993 What do you include in a narrative? A comparison of the written nar-
ratives of Mexican and American fourth and ninth graders.
Pragmatics and Language Learning Monograph Series, Volume 4:
209-224.
Odlin, T.
1989 Language transfer: Cross-linguistic influence in language learning.
Cambridge: Cambridge University Press.
Piaget, J. & B. Inhelder
1956 The child's conception of space. London: Routledge.
Poulisse, N.
1990 The use of compensatory strategies by Dutch learners of English.
Enschede: Sneldruk.
Schiffrin, D.
1987 Discourse markers. Cambridge: Cambridge University Press.
Renee Waara
1. Introduction
In other words, somehow the utterance makes sense, but it sounds odd in a
way that may be attributed to the conceptualization process. Learner con-
structions provide insight into the conceptualization process, which reflects
the state of the learner's language system.
In cases where learners have chosen to use get, they use it in comprehen-
sible ways, and prefer certain constructions to others, but most importantly
they produce slightly off utterances. However, the nature of the task and the
learners' approach to the task may be responsible for the different sense
preferences as well as the choice of verb. In other words, the fact that the
non-native speakers (NNS) use get less frequently than native speakers
(NS) may be related to either avoidance or simply the lack of accessibility
to other linguistic forms. Therefore, what I can say about constructions and
language acquisition is restricted to an analysis of the learner constructions
and how these relate to the actual occurrences of get constructions, the dis-
tribution of these get forms in terms of what is more frequent and which uses
are less frequent, as well as a comparison to the frequency of forms chosen
by the native speakers. The construction senses, e.g., obtain, move, undergo,
that occur in the learner constructions correspond to the most frequent
senses that were used in conventional ways by the learners. The availability
of a construction is an important factor in L2 production because it allows
speakers to take what they have as far as they can, and often this entails that
the extension of a construction overrides the conventional usage. This may
be because conventional usage is less accessible, whereas constructions are
more accessible in that learners are not dependent on knowing or remem-
bering specific linguistic items. In cases where communication is the goal,
constructions allow L2 speakers to be understood simply because construc-
tions capture experience in a language-specific way and already exist in the
Ll system.
Patterns are essential and make our lives easier and livable. The learning
process, in general, is often a matter of seeing a pattern. L2 language learn-
ing may be similar in that in order to grasp an L2 we have to recognize,
remember, and recall the patterns of lexemes, and the combination of these
lexemes, and the inherent semantic connection. Seeing patterns in units that
consist of more than one word has spawned a series of studies on the phe-
nomenon often referred to as "formulaic language". However, describing
patterns and the role of patterns is dependent on the theoretical assumptions
regarding the nature oflanguage.
In the generative linguistics approach (cfChomsky 1965; 1981; 1986;
1995), the focus has been on the relationship between chunks of language
for the purpose of abstract rule formulation to accommodate an approach to
language in which streamlining and high-order generalizations are essential.
Crucially, context and use are viewed as independent of the language system.
Researchers have been particularly interested in the production of "gram-
matically advanced" expressions that appear to be beyond the capacity of a
language user (cfClark 1974; Wong Fillmore 1976; Huang & Hatch 1978;
Wagner-Gough 1978; Hatch, Peck & Wagner-Gough 1979; Vihman 1982;
Karnio11990; Willett 1995; Henry 1996; Myles, Hooper & Mitchelll998;
Myles, Mitchell & Hooper 1999). Many researchers have suggested that a
language learner performs some form of reanalysis of chunk language, result-
ing in the formation of productive rules. However, reanalysis is an assumption
held within the framework of universal grammar, and the problem becomes
making a concrete connection between the use of larger chunks of language
and universal grammar. This has been referred to as the linking problem (cf.
Tomasello 2000: 232ft). If a theory of language stipulates a core grammar
(universal grammar) that is accessible to the language user's a priori lan-
guage input, then the theory must also account for the link between universal
grammar and the particular language learned by the language learner.
Tomasello (2000: 232ft) proposes that this is an inherent problem for a theory
that stipulates a universal grammar. A linking problem would apply equally
to L2 language learners. Consequently, studies on L2 acquisition within the
generative linguistics framework have a linking problem between the param-
eters to be set in the universal grammar and the particular language used by
the language learner.
In a usage-based approach to language, there is no linking problem
because rule formation can only derive from specific, concrete expressions
that are already in the language user's possession. Working within a cognitive
linguistics framework, Dabrowska (2000) has shown how language chunks
are utilized in the formation of schema extraction through a process moving
from lower level schemas to more abstract schemas. The acquisitional process
of emerging schemas at various levels of abstractness is reconcilable with
patterns found in learner language as well as the notion of categorization. If
verbs are learned on a verb-by-verb basis, then repetition of these patterns
will be found in language production in the form of chunks of language or
formulaic language. Language acquisition is then perceived as a process of
categorization of patterns rather than a reanalysis of chunks of language for
abstract rule formation. These chunks are related to constructions in that they
are part of the inventory of language knowledge, and from them constructions
emerge.
(2) sneeze
a. Pat sneezed.
b. Pat sneezed the napkin off the table. (caused motion construction)
c. She sneezed a terrible sneeze. (cognate object construction)
d. She sneezed her nose red. (the resultative construction)
e. She sneezed her way to the emergency room. (the way construction)
Sneeze is a "standard" intransitive verb, but we see from (2b) that it can
combine with three other syntactic components. Goldberg argues that it is
unnatural to assume that sneeze is responsible for the fact that someone
caused something to move somewhere. If we attribute the caused motion
meaning in (2b) to the caused motion construction, then there is no need to
specify a special sense for sneeze. Moreover, the additional senses found in
examples (2c-e) are attributed to their respective constructions. These exam-
ples can be accounted for by the way in which the verb sneeze is construed in
a particular construction. It is not the construction that has changed meaning,
but rather that a situation is construed and expressed in words. The notion
of construction as presented here is also extremely useful for verbs that are
less specific than sneeze, the so called "light verbs", e.g., put, do, make. and
get, in that the form-meaning correspondence is neatly conveyed in con-
structional meanings, as well as the variety and diversity of possible argu-
ment structures. In the next section the corpora from which the data are
extracted are described.
3. Method
The data used in this study are taken from two corpora that were collected
for a speaking test, including a non-native speaker corpus and a native
speaker corpus. The non-native speakers were Norwegian 14--15 year olds,
while the native speakers were 14-15 year olds from schools in the
Newcastle area. The non-native speaker corpus contains approximately
33,000 words (tokens), and the native speaker corpus contains some 20,000
words. The non-native speaker corpus has more individual subjects than the
native speaker corpus, 58 compared with 26. A tester administered every
test in pairs, so there are 29 pairs in the non-native speaker corpus and 13
pairs in the native speaker corpus. Each test took an average of 30 minutes,
of which approximately half is tester talk, so there are approximately 435
minutes of non-native pupil speech and 200 minutes of native pupil speech.
Both the natives and non-natives pupils took the same set of tests. The tasks
are designed to elicit evidence of the various components of communicative
language ability, and presented in detail in Hasselgren (1998).
There are two different versions of the speaking test in the corpus. The
structure ofthe test is the same, with different themes. The theme of version
one is a disco and that of version two a football match. In the non-native
speaker corpus there were 17 pairs doing version one and 12 pairs doing
version two. In the native speaker corpus there were six pairs doing version
one and seven pairs doing version two. The division between the two different
versions has no effect on the analysis because while the themes are different,
the test design ensures similar conditions and comparable contexts for lan-
guage production.
The tasks taken from the corpora represent four different types.
In task one, Pupil A tells the story using the picture series. These are generally
longer stretches of speech, with only minor interruptions from the tester,
almost never from the other pupil. In task two, pupil B is asked to talk for
two minutes on something slhe did recently. This is something unique to the
pupil, no pictures. In task three, one of the pupils explains in detail to the
other pupil how to do something in the pupil's home. There is not usually
very much interaction between the pupils. In task four, one of the pupils
gives directions to the other pupil.
In task one, pupil A is asked to tell a story based on the series of pictures
she has in her test booklet. In version one the theme is a disco. A girl has a
curfew that she has broken, and when she gets home there is a man breaking
into her family's car. In version two, there are two teenagers who are on
their way to a football match, and they miss their train because they left
their bag back in a cafe. They set out to hitchhike and get a lift with the
director, winding up in the director's box. In short, there is enough action
for good storytelling. Pupil A's speech in this section is generally longer
stretches with only acknowledgements andlor encouragement from the
tester. Rarely does the other pupil intercede with a comment. The tester
tends to be most active towards the end of the storytelling, while trying to
tease out the action in each story, with statements from version two such as
where are they sitting? In the picture there is a thought bubble that says
"thief get dad". The speakers tend to put this in rather strange contexts, such
as she saw a thief get dad (the thief didn't get her dad) or she thought/thinks
thief get dad (in which the pupil is telling us what the girl is thinking),
which indicates that they are reading the text. There are more of these in the
NNS than in the NS corpus, and the former seem to be motivated to use the
phrase because they know it is correct language and want to get it in the
conversation.
In task two, pupil B is asked to talk about something she did recently. In
version one, the time frame is set specifically to last weekend or a time she
went out with her friends, and in version two, pupil B is supposed to talk
about a journey or event she attended. There is often a rather lengthy negoti-
ating section before the pupil actually starts her two-minute talk. Each pupil
talks about something that is unique to her; i.e., there are no pictures, and
the listeners are not familiar with what the other pupil is talking about. The
difference between this task and the other tasks is that there are no pictures,
or predetermined guidelines for the pupils to follow.
In task three, one pupil is to instruct the other pupil on how to do a chore
in her/his own home. In version one, the chores include cook potatoes, wash
clothes, make coffee or tea, and record on the video-player. In version two,
one of the pupils has to instruct the other about how to take care of the other
pupil's puppy for the afternoon. There is a detailed picture showing the
kitchen and backyard that is to be used in instruction giving.
Giving directions is in task four. In version one, the task is to retrieve a
CD from the other pupil's house. One pupil directs the other inside herlhis
house to where the CD is located. In version two, one pupil is supposed to
be living in Slode, and is expecting a visit from the other pupil. The pupil in
Slode has to provide explicit directions on how to get to Slode via Newcastle
and Coventry. Detailed timetables and platform numbers are provided.
Language produced in this section occurs in longer stretches, and although
lexical searches are common, they do not tend to break down the communi-
cation.
The speech that is produced by these speakers is not natural conversation,
with the possible exception of the two-minute talk, but even that is guided
through topic choice. Instead of free conversation, we have pupils who are
talking within specific contexts under the duress of the speaking test frame-
work.
Be * *
Go * *
Have
Take * * * *
Do * *
Come *
Get * * * *
Think
See
Put *
Significant difference p < .05 indicated with *
Among the most frequent verbs found in both the corpora, light verbs such
as go, do, get, put are highly represented, thus coinciding with verbs that are
learned and produced in L1 at high frequency levels (cf Clark 1978).
Moreover, light verbs appear as the most frequent verbs in the two corpora.
Within the four tasks selected for this study, the following verbs were the
most frequent. The ten most frequent verbs in order of frequency in the
NNS corpus are listed in table 2.
The verbs that are significantly different are marked with *. Get was chosen
for this paper because it is one of the verbs that differs significantly with
respect to frequency of occurrence in each task. Every occurrence of get
was extracted from the corpora, i.e., get, got, getting, gotten, and to get. For
each sub corpora each occurrence of get was classified by its form-meaning
construction. Table 3 lists the total number of occurrences in each task:
NNSs 34 16 18 11 79
NSs 31 22 40 48 141
Following have and be, get is the most frequent verb form in conversation.
(cf. Biber et al. 1999; Krug 2001) Get is a prime example of a verb with no
clear argument structure because it appears in so many different construc-
tions. This is not a property solely attributed to get; according to Biber et al.
(1999) the next eight most frequent verbs are just as difficult to assign a
clear argument structure, i.e., say, go, know, think, see, come, want and mean.
Hence the more frequent a verb is, the more combinations it tends to occur
in. Thompson & Hopper (2001: 50) take the lack of argument structure a step
further and claim that "The data show very clearly that the most frequent
verbs in the language have no 'argument structures', but occur in a wide
range of lexicalized expressions that must be learned." Thus it would seem
that get combines with many different constructions and therefore it is diffi-
cult to know what kinds of argument structures are permissible. However,
the data in this study at hand indicate a higher degree of regularity.
Get occurs in a variety of combinations from idiomatic instances to sys-
tematic patterning. Idiomatic in the sense that the sum of the parts does not
provide meaning, but rather refers to something completely different; for
example, get it 'understand', get it over with 'suffer through', get into it
'become interested', get the phone 'answer', get down 'party'. Idiomatic
occurrences were extremely rare in the corpora. There were only two
instances in the data: one in NNS in the instruction task, I didn ~ get it,
meaning lack of understanding of what was said; and one in NS, in the two
minute free speech task, as we got into it, meaning 'as we became more
involved' and not into a car, for example. The systematic patterning of get,
in terms of constructions, is described in Table 4.
Table 4 is divided into three columns: construction meaning, basic sense,
and construction form. The form-meaning correspondence is represented in
each row, such that "X OBTAIN Y" will always correspond to a transitive
construction with the basic sense OBTAIN. Table 4 gives the range of
senses of get. However, other verbs are also possible in these patterns, to
some degree. For example, bought can express OBTAIN, as in I bought it at
the store, instead of I got it at the store. In other words, these constructions
bear the meaning of "obtain", "cause obtain", "cause someone to receive
something" so even though other verbs combine with constructions render-
ing a slightly different reading, bought versus got, they do not occur in such
a diversity of constructions. Whereas a verb such as buy can combine in
OBTAIN and CAUSE OBTAIN, get nms the gambit from OBTAIN to DO
as illustrated in Table 4. So in a manner of speaking, get is a "watered-
down-pretty-much-do-anything" kind of verb, which appears to get meaning
through specific constructions, i.e., the combination of the meaning in the
left-hand column of the table and the form in the right-hand side of the table.
For each construction in Table 4 there is a non-causative reading and a
causative reading, with the exception of the basic sense "obligation". Each
syntactic form has a corresponding semantic sense; if there is a change in
syntactic form, then there is a change in meaning. Each occurrence of get in
the corpora has been categorized according to the constructions described in
Table 4.
Given that get can integrate with a variety of constructions, we might
expect that NNS speakers would use it incorrectly. If there are no discernible,
reliable argument structure patterns, then each individual use must be learned
and can only be achieved through contact with many examples. It does,
however, appear that, despite the surfeit of examples, the L2 learners in this
corpus manage quite well. Their success is reflected in the fact that commu-
nication rarely breaks down, and that restarts, which indicate hesitation and
uncertainty, are not overly common. However, the existence of a set of
learner constructions suggests that get (and perhaps light verbs in general)
presents some challenges for learners in terms of knowing precisely which
arguments are permissible or acceptable in particular situations. This uncer-
tainty is realized by the emergence of conceptual blends, elements of transfer,
and overgeneralizations.
4. Results
The frequency of occurrences differs significantly between the NNS and the
NS speakers. In Table 5, all of the occurrences of get are categorized in
terms of basic sense or other combinations that emerged from the data.!
The emergent combinations are four non-sense categories that are
included to highlight other aspects of the uses of get. For example, when get
occurs in the infinitive form as in trying to get the car, even though the basic
sense of OBTAIN is obvious, there is another verbal element that flavors its
OBTAIN 29.4 19.4 12.5 22.7 16.7 65.0 9.1 18.8 20.3 32.6
C-OBTAIN 2.9 0 0 4.5 0 2.5 0 0 1.3 1.4
MOVE 14.7 45.2 37.5 18.2 0 7.5 81.8 41.7 25.3 29.1
C-MOVE 0 6.5 0 13.6 5.6 0 0 22.9 1.3 11.3
BECOME 5.9 0 6.3 13.6 5.6 0 0 0 5.1 2.1
C-BECOME 2.9 0 0 0 0 7.5 0 2.1 1.3 2.8
UNDERGO 0 3.2 0 4.5 0 0 0 2.1 0 2.1
C-UNDERGO 0 0 0 0 0 2.5 0 2.1 0 1.4
DO 0 0 0 0 5.6 0 0 0 1.3 0
C-DO 0 0 0 0 0 0 0 0 0 0
OBLIGATION 0 9.7 0 0 5.6 10 0 4.2 1.3 6.4
INF+OBJ/OBL 26.5 16.1 12.5 13.6 0 2.5 9.1 2.1 15.2 7.1
RESTART 5.9 0 6.3 4.5 5.6 2.5 0 4.2 5.1 2.8
IDIOMATIC 0 0 0 4.5 5.6 0 0 0 1.3 0.7
LEARNER 11.8 0 25 0 50.0 0 0 0 21.5 0
tions, which represent 52 percent in the NNS and 70.2 percent in the NS.
The NNS and NS are both relatively low for the causative and relatively high
for the non-causative, showing similar tendencies. In order to illustrate how
the uses of get are realized in the corpora, below is an example of each con-
struction:
OBTAINIPOSSESS: you get the dogs dish/ she 50 got the bag
- C- OBTAIN: you'll have to get it a drink
- MOVE: they get to the railway station
C-MOVE: we'll be getting the next train
BECOME: it 50 got dirty
C-BECOME: they may not get them right
UNDERGO: you'll get killed
C-UNDERGO: its got "dog" written on it
DO: so he'll get to know you
OBLIGATION: you've got to go upstairs
INF + OBJ/OBL: she went to get her dad
RESTART: when we got when here we come to the
- IDIOMATIC: I didn't get that
LEARNER: she get to know that she must be home
If the focus of study is solely on the combination of words with a strict top-
down abstract set of rules or principles of language, then learner constructions
are quite difficult to account for because syntactically they are generally
correct. However, from a constructional perspective we can assume that
underlying knowledge that verbs have different senses influenced the
speaker's choice. Activation of possibilities must play an important role, a
certain degree of automaticity with the pressure of the online processing.
Learner constructions sound slightly odd because they rely more on the
application of a constructional sense than a conventional expression. In a
general way, constructions dominate at the expense of conventionality.
Specifically, with the notions of construction as the point of departure,
learners blend conventional uses within the target language, transfer aspects
ofLI constructions into the L2 system, and overgeneralize existing L2 con-
structions in their developing L2 system.
The overall results will be discussed first. Then, each sub corpora which
correspond to the four individual tasks described in section 3.1 will be dis-
cussed in terms of group tendencies and preferences for certain senses of
get. Finally, learner constructions will be discussed in terms of how they
This task involves telling a story based on a series of pictures and was dis-
cussed in more detail in section 3.1. In terms of construction distribution,
when they are telling a story the NS clearly prefer the MOVE sense, as in
they get to the disco, and this represents 45 percent of the total number of
occurrences in task one. In contrast, the NNS do not utilize the MOVE
sense as frequently as the NS, and prefer the categories OBTAIN and the
INF+XP. In addition, the NNS produce the most instances of get in task
one, i.e., 43 percent of all the instances are found in task one. The distribu-
tion suggests a different approach to the task, such that instead of telling a
story the NNS speakers describe the pictures, such as she s got the bag,
whereas the NS speakers utilize the MOVE sense, such as they get to the
railway station, in accordance to what we would anticipate in telling a story,
i.e., the movement of a narrative.
In task two, pupils are asked to talk about a recent event or journey. The dif-
ference between this task and the other tasks is that there are no pictures or
predetermined guidelines for the pupils to follow. The pupils introduce a
variety of experiences from babysitting, football games, and shopping. Thus
the topics have no commonality apart from the fact that they are not imposed
as a specific task. We might expect this choice to allow the pupils freedom
to avoid language they are uncertain about. This is also the task in which
both the NS and the NNS produce the least number of occurrences of get. In
contrast to task one, the usage differs. Whereas the NNS strongly favor the
MOVE sense, e.g., I got home and We got in, the NS favor OBTAIN, e.g.,
I've got money. The use of the MOVE sense is an indication that this con-
struction is accessible to the NNS speakers when they see reason to express
movement.
In task three, one pupil instructs the other pupil on how to do something.
Thus we would expect to find a sequence of actions, most likely connected
with and then. This is what we find. The NS produce most OBTAIN con-
structions, e.g., you get the dog s dish, and it s got a plug. There is a skewing
effect towards possessive construction caused by the fact that the native
speakers were told to pretend they did not know certain words. The project
leader responsible for the project was uying to simulate L2 conditions for
the search for lexical items. Consequently there are several instances of
descriptive strings, e.g., its got holes/ dials/ circles, which complicates the
data interpretation. However, there is another rather interesting result from
this task, in that there appears to be a structural difference that may be related
to a cultural difference. The UK pupils have a slightly different approach to
the matter of telling somebody what to do. Following the and then signal,
the UK pupils prefer get + object, as in and then you get the tin/the towel/
the dogs dish/dirty clothes, whereas the Ll Norwegian speakers prefer
combinations such as and then you got to put/search/looklfill. The UK
pupils focus on things, whereas the Norwegian pupils focus on actions in a
way that breaks with a convention of politeness when instructing someone
to do something.
In task four, one pupil gives directions to a house or place to another pupil.
Of the four tasks discussed here, there are the most occurrences of get in the
NS and the least in the NNS, 34 and 13.9 percent respectively. There is a
strong preference or tendency for MOVE in the NNS corpus, e.g., get off.
There is a clear preference in the NS corpus towards the MOVE and the
CAUSE MOVE constructions, e.g., we 'U be getting the next train to
Coventry. The NS speakers take the MOVE construction a step beyond with
the causative reading, and specify both movement and destination. 2
In this section the emergence and usage of the learner construction will be
discussed in terms of how they reflect elements of transfer, blending, and
overgeneralizations of the developing L2 system. The intentions of the
speakers in these corpora are identifiable insofar as they are related to specific
tasks, which provide a starting point to distinguish how the speakers construe
a given expression. In other words, viewing the utterance in light of the sur-
rounding context allows us a clearer picture of the intentions of the speaker.
The learner constructions are divided by tasks, i.e., the examples in (a-d)
are from task one, the examples in (e,f) are from task two, and the examples
in (g-j) are from task three. No learner constructions occurred in task four.
Learner constructions in task one:
In (a), the form is subject-verb-object, dance is an object and suits the transi-
tive form, but a dance requires two people, albeit under traditional conditions.
Can I get has the conventional use of offering to get something for someone,
as in can 1 get you a coffee? The conventional expression used to obtain a
dance with someone is May I have this dance? The example in (a) gives us
a blend of two conventional usages instantiated by can I OBTAIN a dance
with you?
In (b), the MOVE construction captures the sense that the man is moving
towards and arriving at the car. The speaker expresses or describes what is
in a picture, a masked man with his hand on the car handle. Based on the
series of pictures we can tell that he has just moved over to the car. The
speaker has used the intransitive motion construction to describe this event.
The man moves along some path to the goal, which is the car. When get is
used in the intransitive motion construction it has restricted conventional
usage. These uses include expressing future events, e.g., when we get to the
cabin, or past-completed action, as in we got home late. The example in (b),
a "man get to the car", is used in an unconventional manner to express "X
moves Y" because it overgeneralizes the use of get within the speaker's sys-
tem of L2 language.
In (c), the thief is undergoing a process. In the picture, the thief is literally
being shaken by a man, so this construction is the passive construction, "X
undergoes Y". It would seem that the meaning of getting is being extended
to the meaning of being, and the construction contributes most of the mean-
ing. Moreover, in the two-minute task there is a NS example of, we were
gettingfollowed. In other words, the form is acceptable and it is not only the
NNS who can stretch a construction, i.e., constructions have the ability to
be stretched.
In (d), the parents are telling the girl that she has to be home by midnight.
The meaning is retrievable from the context; she OBTAINS, in a loose sense,
a message from her parents, but the form is confusing because get + to-
infinitive verb form denotes some action in "X acts on Y", and yet the
intended meaning of the expression is passive, in the sense that the girl is
told something. However, there is a transfer effect from the Norwegian, afa
vile noe 'to get to know something', which has the basic sense 'receive' or
'obtain'. But we also 'get to know someone' which conventionally refers to
'be acquainted with someone'. This learner construction is a transfer from
L 1 combined with an over generalization with the existing L2 system.
There are three very different learner constructions found in the NNS
task two.
rendition or use ofla 'get', i.e.,flkk problem med en annen 'got problem
with one other'. This is a case of overgeneralization of the basic sense of
OBTAIN and L1 transfer.
There are not many direct transfers from Norwegian; however, the
example in (t) is a clear exception, i.e., du blir bedre kjent med noen 'you
become better known with someone'. The construction is very descriptive
in that "X becomes Y". Get relates the subject to the state of 'getting to
know somebody'. Although this is an unconventional use, communication
is successful. The Norwegian use of get in (e) and in (t) instantiates Ll
transfer, which may be related to the strong relational sense between
become and the Norwegian verb blir.
In task three, the following learner constructions are found:
The examples in (g) correspond to the construction meaning 'X acts on Y'
and is, on its own, an acceptable expression. However, in the course of giv-
ing instructions, it is generally not conventional to imply that someone must
do something, because of the edge of impoliteness it implies and because it
diverges from a general convention of politeness in the given context.
In (h), the aroma is attributed with a property of motion; the aroma moves
away to someplace else or is "dispersed". The pupil has described how to
boil the water and put the coffee granules in the water and let it stand. The
pupil says that for filter coffee the powder is ground more finely than for
boiled coffee, and because "the aroma gets sooner away", he prefers the
boiled coffee. In other words, the flavor of boiled coffee is better because it
does not lose the aroma by being ground as finely as the filter coffee. The
verb disappear is suggested by the tester in this exchange, and even used by
the pupil earlier in the turn, but for some reason it is later rejected and a
conceptualization process involving "X moves Y" is preferred by the pupil.
The pupil feels more expressive or more accurate by describing the aroma
in a MOVE construction, based on a basic human experience, "move some-
thing". This example is an overgeneralization within the existing L2 system.
The example in (i) is conventional in the sense that you can say, Get
upstairs!, with the reading 'go or move upstairs'. But in this context it is
slightly odd because it is a part of a series of directions about moving from
one point to another point, i.e. first you go there and then you go there. In
5. Conclusion
What is striking is that it looks like the learners have mastered the construc-
tion level better than specific lexical items, such that the construction is
supplied with semantic components that are either 100% compatible or
slightly unconventional, hence giving us a learner construction. The problem
or point of conflict is the semantic compatibility of the components in the
construction. Although many of the get constructions produced by the learn-
ers are acceptable and in many cases quite similar to the native speaker
choices, there are some differences found in the learner constructions, dif-
ferences that suggest that constructional meaning guides production at the
expense of complete conventional usage. Perhaps these speakers do not
have the lexicalized expressions in their language inventory. Evidence for
this is found in the overgeneralization of constructions and descriptive
power of the construction.
In the majority of the learner constructions presented here, the construc-
tion itself is a composite whole. It is the interaction between construction
and its components in a specific context that produces the unconventional
uses. That is, the manner in which the speaker construes the words, i.e.,
Notes
1. Cases of 'thief get dad' were not included because it is assumed to be influenced
by the language bubble in the picture that the subjects used to tell the story.
2. It is worth noting that both English and Norwegian are satellite-framed languages
in Talmy's (1991) typological framework of motion events.
References
Achard, Michel
1997 Cognitive Grammar and SLA Investigation. Journal of Intensive
English Studies 1112: 157-176.
Biber, Douglas, Stig Johansson, Geoffrey Leech, Edward Finegan & Susan Conrad
1999 Longman Grammar of Spoken and Written English. London:
Longman.
Chomsky, Noam
1965 Aspects of the theory ofsyntax. Cambridge, MA: MIT Press.
1981 Lectures on government and binding. Dordrecht: Foris.
1986 Knowledge oflanguage: lts nature, origin, and use. New York: Praeger
Publishers.
1995 The Minimalist Program. Cambridge,MA.: MIT Press.
Clark, Eve V.
1978 Discovering what words can doPapers from the Parasession on the
Lexicon, Chicago Linguistic Society, 34-57.
Clark, Ruth
1974 Performing without competence. Journal of Child Language 111: 1-10.
Dabrowska, Ewa
2000 From formula to schema: The acquisition of English questions.
Cognitive Linguistics 11/112: 83-102.
Fauconnier, Gilles
1997 Mappings in thought and language. Cambridge: Cambridge University
Press.
Goldberg, A. E.
1995 Constructions: A Construction Grammar Approach to Argument
Structure. Chicago: University of Chicago Press.
Goldberg, Adele.
1998 Patterns of Experience in Patterns of Language. In M. Tomasello
(ed). The New Psychology of Language. New Jersey: Lawrence
Erlbaum Associates, 203-219.
Hasselgren, Angela
1998 Smallwords and Valid Testing. Ph.d, University ofBergen.
Hatch, Evelyn, Sabrina Peck & Judy Wagner-Gough
1979 A Look at Process in Child Second-Language Acquisition. In E. Ochs
& B.B. Schieffelin (eds). Developmental Pragmatics. New York:
Academic Press, 269-278.
Henry, Alex.
1996 Natural Chunks of Language: Teaching Speech Through Speech.
Englishfor Specific Purposes 1511: 295-309.
Huang, H & Evelyn Hatch
1978 A Chinese Child's Acquisition of English. In E. Hatch (ed). Second
Language Acquisition, A Book of Readings. Rowely, MA: Newbury
House, 118-131.
Karniol, Rachel
1990 Second-language acquisition via immersion in daycare. Journal of
Child Language 17: 147-170.
Krug, Manfred G.
2001 Frequency, iconicity, categOl;zation: Evidence from emerging modals.
In J. Bybee & PJ. Hopper (eds). Frequency and the emergence oflin-
guistic stntcture. Philadelphia: John Benjamins, 309-336.
Langacker, Ronald W.
2000 A Dynamic Usage-Based Model. In M. Barlow & S. Kemmer (eds).
Usage Based Models of Language. Stanford, California: Center for
the Study of Language and Information (CSLI).
Myles, F., 1 Hooper & R. Mitchell
1998 Rote or rule? Exploring the role of formulaic language in classroom
foreign language learning. Language Learning 48/3: 323-364.
Myles, Florence, Rosamond Mitchell & Janet Hooper
1999 Intenogative Chunks in French L2: A Basis for Creative Construction?
Studies in Second Language Acquisition 21/1: 49-80.
Pawley, Andrew & Frances Hodgetts Syder
1983 Two puzzles for linguistic theory: nativelike selection and nativelike
fluency. In le. Richards & R.W. Schmidt (eds). Language and Com-
munication. London: Longman, 191-223.
Talmy, Leonard
1991 Path to realization: A typology of event conflation. In Teh Seventh
Annual Meeting of the Berkeley Linguistics Socie~v: 480-519.
Berkeley: Berkeley Linguistic Society.
Thompson, Sandra & Paul 1 Hopper
2001 Transitivity, Clause, and Argument Structure. In J. Bybee & P.J.
Hopper (eds). Frequency and the emergence of linguistic stntcture.
Amsterdam: John Benjamins Publishing Company, 27-60.
Tomasello, Michael
2000 Do young children have adult syntactic competence? Cognition:
internationaljournal of cognitive psychology 74/3: 209-253.
Vihman, Marilyn M.
1982 Formulas in First and Second Language Acquisition. In L. Menn &
L. Obler (eds). Exceptional Language and Linguistics. New York:
New York Academic Press, 261-284.
Wagner-Gough, Judy
1978 Excerpts fonn comparative studies in second language learning. In
E.M. Hatch (ed). Second Language Acquisition: a book o{readings.
Rowely,MA: Newbury House, 155-174.
Willett, Jeni
1995 Becoming First Graders in an L2: An Ethnographic Study of L2
Socialization. TESOL Quarterly 29/3: 473-503.
Wong Fillmore, Lily
1976 The Second Time Around: Cognitive and Social Strategies in Second
Language Acquisition. Ph.D., Stanford University.
1. Introduction
• 'E
Lemma Seman:ic / ~
Lexeme node
FO~. ~
·~~I if
• U§
Syntactic
properties
This image of lexical item in the mental lexicon can be adjusted to the bilin-
guallexicon by assuming an additional source of infonnation linked to the
lemma node referring to the language a lexical item is associated with. The
items associated with a particular language can be regarded as a subset of
the lexicon (cf. Lowie 1998; Woutersen 1997).
A question that is relevant to the bilingual mental lexicon is whether and
to what extent L2 learners make use of the lexical knowledge from their
first language in the acquisition and use of the second language. As the
adult L2 learner has a fully developed lexicon, it makes sense to assume that
an L2 learner will make use of the knowledge already required. A question
that is central in current debates on the bilingual mental lexicon is whether
L2 words have direct links with conceptual memory or are accessed through
Ll lemmas present in the lexicon. Recently, Nan Jiang (2000) argued that
the role of the first language differs in three stages of development. In the
first stage, L2 fonns are mapped onto existing (Ll) meanings. At this stage,
an "empty" L2 lemma is created that is linked to a Ll lemma: the L2 lexical
item only has the fonnal characteristics and full equivalence to an L 1 lexical
item is assumed and there is no direct link from the L2 lemma to conceptual
content. At the second stage, the infonnation of an existing Ll lemma is
copied onto the L2 lemma: this is the situation where the Ll lemma mediates
L2 word processing; now there is a link from the L2 lemma to both concep-
tual content and the Ll lemma. Only at the third stage will the Ll lemma no
longer be accessed and a direct link has been created between the conceptual
representation and the L2 lemma. Jiang's model sketches a picture similar
to the one proposed by, for instance, Kroll (1993), in which lexical items in
L 1 and L2 are connected:
Although this model conveniently explains what Jiang calls "lexical fos-
silization", it cannot account for the fact that lexical items in Ll and L2
hardly ever fully overlap in meaning. In a model that takes compositional
meanings as a starting point, this can be accounted for much more easily.
By referring to the activation metaphor, it is no longer necessary to distin-
guish between different ways of lexical organization; activation models
hypothesize that all individual lexical entries are stored identically, but that
major differences between the entries can be expected based on their fre-
quency, expressed by their relative level of activation. Ll entries are never
directly linked to L2 entries, but information shared between the languages
will result in activation feedback flowing to the lemma nodes concerned. In
other words, Ll and L2 entries can never be lexically mediated, but are
always conceptually mediated to a degree dependent on the relative activa-
tion of the conceptual representations, the lemma nodes, and the lexemes.
Similar to how partially overlapping meanings in the monolingual mental
lexicon can be accounted for, the activation model can also account for over-
lapping meanings between Ll and L2. Figure 4 exemplifies the partial over-
lap between a Dutch and an English item in the bilingual mental lexicon.
The same framework can also be used to account for the development of
the bilingual mental lexicon. At initial stages of L2 acquisition, a full overlap
may be assumed between the conceptual representations of the Ll lemma
and the L2 lemma. Gradually, the differences between the L1 and the L2
lemma will be acquired, which may eventually lead to a "native-like" lexical
representation. This process can be entirely based on positive evidence and
is guided by the same principle of contrast that is at work in L1 acquisition
(cf. Clark 1993). When the learner encounters a new L2 word, this may lead
to the partial restructuring of the semantic form of existing concepts by
adding or deleting the match with some of the conceptual representations.
This process is exemplified in Figure 5. 2 At some early stage of acquisition
(t 1), the Dutch learner of English will assume full overlap between between
and among, since Dutch does not make this conceptual distinction. 3
Subsequently, the principle of contrast will ensure that the learner will not
accept two fully identical lemmas, leading to the discovery of the semantic
differences between between and among. This will then lead to restructuring
of the semantic form of between and the creation of a new lexical item
among. The ultimate result of the acquisition process can be a "balanced"
bilingual lexicon in which all semantic forms of all lemmas have been fully
specified. However, cases where this happens for all lexical entries in both
languages will be highly exceptional, as most bilinguals will not be fully
"balanced". The additional advantage of this approach is that it is no longer
necessary to assume the same stage of development for entire language sub-
sets. While some L2 lexical items may be fully developed, including all
semantic and syntactic regularities and restrictions, other items may be found
in different stages of acquisition.
0
0 0
eerlijk o O~o DUTCH
D----o~g
HONEST
FAIR
o 0 ADJ
o 0 +ABSTR
0 - - - ~O
fair
o 0 ENGLISH
0 0
0
0 0
LX SF CR LX SF CR
o o
000
00 0 among
tussen o
0-1,40 o~~ 0 tussen
00 0
_-0
between 0-1,40 0 between
o 0 o
o
Figure 5. Simplified representation of two time slices (tl and t2) in the process of
acquiring the new L2 concept "among". In this figure, the lemma nodes
have been left out. LX= lexeme; SF = semantic form; CR=conceptual
representations.
The other variable factor, frequency, is independent of the mother tongue and
expresses the extent to which the L2 lemma is used. The role offrequency
in activation models is obvious, as processing in the lexicon is driven by
frequency-induced activation; all elements in the lexicon can attain variable
degrees of activation, which increases each time a node is used, and decreases
over time. Activated nodes spread activation to nodes with which they are
connected. Frequency is thus the major drive behind lexical acquisition.
The two factors described here, input frequency representing the inde-
pendent effect of L2 on the acquisition process, and formal similarity repre-
senting cross-linguistic influence, were also the main variables in a recent
Table 1. Overview of the English prepositions and their Dutch translation equiva-
lents in the four categories selected.
described above, our expectations were that, especially at the lower levels
of acquisition, a high degree of similarity between the Dutch and English
prepositions would positively affect the correctness of the answers. The
effect of frequency, which was previously shown to occur only after pro-
longed exposure to the second language, was expected to be strongest at
higher levels of acquisition.
4. Results
The scores of the four groups in this experiment were analyzed using a
MANOVA, with group as the between-subjects variable (four levels) and
similarity and frequency as within-subject variables (two levels each).
The main effect of the between-subjects variable, group, was significant
(F[3,71]= 45.2; p<O.Ol); the highest number of correct scores was found in
the group with the highest proficiency (see Figure 6). Also the main effects
of the within-subjects variables similarity (F[1,71]=71.0; p<O.Ol) and fre-
quency (F[1,71]=82.S; p<O.Ol) were both significant.
100
90T-----------------------------------r=====~--
.. -:-:..-:-.71-,------1
80------·----------·----------r..:-:-.-:-:
70 --------------------------4
60 - j - - - - - - - - - - - - - - - - - j . . . . . . . . . t - - - - - l ..... -
. . . . .. .. .
50 +----------------i t - - - - - l ......... r------j ....... - -
40+----------------i ...... -
30 r-----: .... -
20 r-----.' .,-
O+-~----~--~~----~--~~----~--~~-----J__.
group 1 group 2 group 3 group 4
Also, the interaction between group and similarity turned out to be signifi-
cant ([F3,71]=S.O; p<O.Ol). The analysis showed that for similarity also the
smallest effect was found at the highest level of proficiency (see Figure 8).
The interaction between frequency and similarity (F[1,71]=3S.3; p<O.Ol)
showed that the largest similarity effect was found with the low-frequency
prepositions (see Figure 9).
The three-way interaction between group, similarity, and frequency was
significant as well (F[3,71]=5.1; p<O.Ol). The analysis showed that the inter-
action between similarity (as represented in Figure 7) was only significant for
the low and the intermediate level of proficiency, but not for the highest level.
A qualitative items analysis revealed that the lowest overall scores were
found with English prepositions that indicate a conceptual distinction that
does not occur in Dutch. Examples are the English prepositions over and
above, both represented by boven in Dutch, and among and between, both
represented by tussen in Dutch. The same effect, but less strong, was found
100.---------------------------------------.
__ ---i
60 '~.~~.~~.-
•••••••••••••• - •••••••••• - ••••••••• - •••••• - ••••••- _ . . . .•••••
group
40
_._._._+
Imavo
.~
.....................
~ 20 3vwo
en
eo
1;l
------.
Juni
~o 3uni •
u O+-------------------------------------~
low high
FREQUENCY
Table 2. Representation of correct scores (%) for some individual items in the three
groups.
1 30 lO 40 50
2 12 4 65 65
3 80 29 80 86
for from and of, both represented by van in Dutch. The number of correct
scores for these words was particularly low at the lower levels of English
proficiency. The leamers typically opted for the more frequent altemative in
these cases. Another striking result was that some items that are very similar
in both form and meaning in English and Dutch (e.g., by - bij, as in 'sit by
the fire') had comparatively low scores at the lower proficiency levels.
100,-----------------------------------------,
--~
-- -- -- -- -- -- ---
80
--- --- ---
4 ---
--- --- .......................
...............................................•
..........
..........
60 " .................... .
group
_._._.-+
Imavo
40 .....................
3vwo
------.
luni
20+---------------------------------------~ 3uni •
low high
SIMILARITY
100,-----------------------------------------,
90
80
70 '11""
60
SIM
50
..................•
1
u
~
8 40+---------------------------------------~ 2
•
low high
FREQUENCY
Figure 9. The interaction between frequency and similarity. High and low frequency
are represented horizontally; similarity is represented by the different lines:
the dotted line for low similarity and the continuous line for high similarity.
5. Discussion
The experiment showed a clear effect for both similarity and frequency for
the low and intermediate levels of proficiency, but hardly any effect for the
highest level of proficiency. Considering the high scores of the high profi-
ciency group (see Figure 6), this observation could be attributed to a ceiling
effect for the highest level of proficiency. Apparently the subjects in this
group had acquired full lexical representations for all the prepositions in the
experiment, regardless of degree of similarity to Dutch prepositions or the
frequency in the input.
At beginning and intermediate stages, both frequency and similarity tend
to affect the score. The effect of similarity was in agreement with our expec-
tations and cOlToborates results from previous studies. Apparently, the formal
6. Conclusion
for the main effects of significance and frequency, this is no serious draw-
back as these were within-subjects variables. And although the cross-
sectional design with these groups cannot be considered identical to real
development as measured in a longitudinal design, the clear difference
between the cross-sections taken gives a strong indication in the direction of
the development observed.
This study suggests that up to the intermediate level of acquisition learners
rely on formal similarities with their L1 in using L2 prepositions. Further
research will be conducted to determine whether the translation equivalence
of lexical items, i.e., the amount of conceptual overlap between L1 and L2
lexical items, is noticed if it does not coincide with formal similarity. It can be
assumed that the greatest difficulty for learners can be found in cases where
formal similarity does not coincide with translation equivalence. This cate-
gory (of "false friends") needs further investigation with learners at different
stages of L2 acquisition. Another interesting observation that merits further
investigation is that qualtitative items analysis of the current experiment
revealed that English prepositions that indicate a distinction that does not
occur in Dutch get the lowest scores (e.g., tussen = between / among; boven
= over / above). The difference between the groups for the scores on these
items seems to support the gradual restructuring of L2 lemmas from assumed
conceptual overlap with similar Ll items to independent (though partially
overlapping) L2 lemmas.
The experiment described here is by no means sufficient to support (or
falsity) the entire model of the bilingual mental lexicon described here. Only
one aspect of the model has been tested and many other questions are yet to
be answered. However, with this experiment we hope to have contributed to
the gradually emerging picture of the bilingual mental lexicon that can only
be established by the constant interaction between empirical studies and
theoretical modeling.
The quiz contained a blank for the underlined preposition. The Dutch equiv-
alent was not provided (but is given here to show level of similarity).
Notes
1. However, Kellerman (1987) has pointed out that Dutch learners resist transferring
non-prototypical uses of such words.
2. Analogous to an example worked out in Schreuder & Baayen (1995).
3. Even though the distinction between henveen and among is one that is not
observed by all native speakers much anymore, it still exists. Benveen refers to
two entities; amollg to three or more.
References
Tomasello, M.
First steps toward a usage-based theory of language acquisition.
Cognitive Linguistics 11, 112,61-82.
Woutersen, M.
1997 Bilingual word perception. Ph.D. Thesis, University of Nijmegen.
Nijmegen: Katholieke Universiteit Nijmegen.
Susanne Niemeier
1. Introduction
The purpose of this contribution is not so much to provide evidence for the
Whorf theory complex of Linguistic Relativity itself, but rather to discuss
how the renewed attention on relativity in language and culture may be rele-
vant to current tendencies in foreign language teaching methodology.
However, let us state right from the beginning that relativism as such is
independent from the concept of a first or second language, but applies lan-
guage-internally, and is not used here as a cover term for "cultural stereo-
types".
Over the last few years we have witnessed a very noticeable rise of interest
in Whorf's thoughts and ideas (cf. Lucy 1992 and 1997, Gumperz and
Levinson (eds.) 1996, Lee 1997, Niemeier and Dirven (eds.) 2000, Piitz and
Verspoor (eds.) 2000). This reappraisal is not primarily due to Whorf's cen-
tenary in 1997, but is also based on more solid grounds such as a strength-
ened acceptance as well as a re-interpretation and re-evaluation of Whorf's
scientific research. This renewed interest stems from various sources: First of
all, many scientific domains slowly abandon the idea that knowledge can be
compartmentalized but argue that we have to take a more holistic perspec-
tive. Second, in linguistics, there is no longer any agreement that language
is to be seen as an autonomous faculty - as the generative paradigm had it -
but there is also the view that it is directly interacting with other mental fac-
ulties. For example, the slow but steady rise of Cognitive Linguistics from
the 1980s onwards has from its very beginning emphasized the fact that lan-
guage, culture, and thought are inextricably intertwined, and that therefore
it is not possible to analyze anyone of them without taking the others into
consideration. Much of our linguistic behavior can be shown to be based on
extralinguistic experiences and cultural knowledge. Even if today Cognitive
Linguistics is not as enthusiastic about Whorf's ideas as it seemed to be in its
beginnings - because his ideas do not seem to allow for universalist tenden-
cies - the interrelation of both research directions is nevertheless certainly a
given.
one's own language and culture, and thus potentially generates a different
worldview. All in all, the insights connected to the concept of language
awareness seem to fit well into Whorf's thoughts on the interconnectedness
of language, thought, and culture.
Quite a similar connection can be discovered when reflecting on the
learning target of intercultural competence (cf. Buttjes and Byram, eds.
1991), a possible means to remedy the learners' lack of familiarity with or
even lack of knowledge about the target culture, which until today is mainly
being taught by texts or, less frequently, by audio and video materials. In a
cognitive-linguistic view, culture is present in the language itself, on all its
levels, and can be detected by language work and language analysis.
"Culture" does not reside somewhere outside of language, but is implicit in
its every word and every grammatical construction and is thus constantly
communicated by the language. Detecting and analyzing intercultural
differences via the foreign language can even provide learners with a differ-
ent, more distanced view of their own culture and thus initiate very helpful
processes of a change in their self-image and of more empathy toward the
foreign culture or foreign cultures in general.
As to the learning objective of autonomous learning (cf. Wolff 1994,
Little 1994), this is rather close to the theory of constructivism in language
learning (Wendt 1996). Learners do not necessarily learn what they are
taught, but they construct their own knowledge little by little from the input
that they are offered. They form hypotheses and try them out, either veri-
fying or falsifying them, and in the latter case they form new hypotheses
(this idea is also quite close to the concept of "interlanguage", cf. Selinker
1992). In this way, learners learn rather independently at their own pace and
the teacher's role becomes that of a monitor and of an input provider.
Autonomous learning can thrive best when the learners are offered authen-
tic learning situations (which involve insights into foreign thinking patterns
and into the foreign culture) in a classroom environment where the above-
mentioned principles of language awareness and intercultural competence
are present as well. Of course, most of the language input still has to come
from the teacher, who at the same time shows the learners how to learn and
how to reflect on their own learning strategies. In this way, learners may
construct their own image of the foreign language, the foreign way of think-
ing, and the foreign culture. The connection to Whorfs ideas is not as obvi-
ous here as it was concerning the before-mentioned trends, but it is present
nevertheless as the hypotheses that learners are able to form are necessarily
influenced by the learners' own linguistic and cultural experiences, be it in
their native language or in a foreign language.
Learners are not a homogeneous group; everybody has their own ways and
strategies of how to construct knowledge. This is why teachers should be
aware of the fact that they have to offer different kinds of information to
their learners. Concentrating on multi-channel learning (cf. Ellis 1994)
means that teachers have to cater to various types of learners by, if possible,
offering for example visual, auditory, audiovisual, tactile, verbal-abstract,
and action-oriented approaches to the information they want to transmit. In
this way, knowledge gets stored in various parts of the brain and different
brain links are constructed, which again means an easier retrieval of the
stored knowledge (see also Lamb 2000). A related approach is that of holistic
learning and teaching (cf. Timm 1995, Loffier 1996) which claims that
cognitive and affective aspects should go hand in hand (involving both
hemispheres of the brain, where traditional teaching was mainly centered on
the right hemisphere) and where all the learners' senses should be activated.
This is supposed to heighten the learners' motivation as well as their atten-
tion and creativity, and to have them enjoy learning (about) the foreign lan-
guage and the foreign culture. Whorf' s ideas can be called holistic as well,
as they highlight the connections between language, thought, and culture
and thus do not see them as separate modules.
Although none of these trends in EFL teaching methodology can be
explicitly connected to Whorf's ideas, they are still in line with the idea that
language is only one facet of a person or a culture, and that it is in a very
close relationship with other mental capacities. Activating these other capac-
ities helps the process of learning and of discovering and understanding
connections, for example those between language and culture, because
culture is not only expressed language-internally, but also, as cultural anthro-
pology continues to point out (cf. Palmer 1996), in other spheres oflife, be it
housing, dressing, agriculture, kinship systems, and the like. These aspects
have to be taken together, and this is not possible by using only language
but needs other channels and other ways of making connections; for example
by raising the learners' interest and emotional involvement, which again
forms part of a holistic way of teaching.
To give another example, the action-oriented approach to learning (cf. Bach
and Timm 1996) fits nicely into the pattern outlined above as well. According
to this approach, learners are expected to develop a competence for acting in
the foreign language, first in the classroom but ultimately also with foreigners
in their own country or within the foreign language culture. The teaching
relies heavily on the methodological approach of "learning by doing", which
involves offering authentic learning environments (as, for example, project
work) and pursues the goal that the learners relate affectively as well as
cognitively to the learning tasks. The learners are expected to use the lan-
guage as much as possible and not just to passively take it in, and by using the
language they learn about its culture-relatedness at the same time. The teacher
has to tolerate a higher rate of mistakes because the general motto is "message
before accuracy", giving the learners the possibility of taking communicative
risks (as we all do in everyday discourse), of building their own hypotheses
about language in general and the foreign language in particular, and of en-
hancing their linguistic as well as their social skills in communication.
Last but not least we should mention the rather recent focus-on-form
approach. Although communicative language teaching (CLT) is probably still
the most widespread view on foreign language teaching, we are currently
witnessing a movement back from a concentration on mere communicative
competence to a so-called "focus-on-form", aiming at reintegrating grammar
into the foreign language classroom by combining formal instruction and
communicative language use (cf. Doughty & Williams 1998). SLA re-
searchers such as Pica see "the need for direct instruction and corrective
feedback" (2000: 11). Another recent development is discourse-based ap-
proaches to grammar instruction, which focus on authentic language uses and
structures and their meanings in discourse and text, i.e. an application of
corpus-based research (cf. McCarthy 1991), so that learners may be con-
fronted with "real" language and not with specially constructed texts. Again,
if we take a cognitive-linguistic perspective, we can claim that grammar is
as closely connected to culture as every other aspect of the language, and
that therefore a renewed focus on grammatical structures, combined with
other approaches, helps raise (inter-) cultural awareness.
It is striking that in all these newer approaches within foreign language
teaching methodology, culture is seen as an inherent part of language and
that every single approach, as different as they may each be, aims at inte-
grating learning the language with learning about the language as well as
with learning about the culture. The main idea, then, is that foreign language
learning or learning a language different from one's mother tongue also
opens up the possibility of seeing the relativity of one's own language and
culture. Language awareness raising combined with foreign language
awareness raising creates the potential of freeing oneself from one's own
prejudiced Ll views. This is not merely a linguistic aim but also a peda-
gogical one. Taking this perspective also aims at developing the learners'
personalities by offering them the possibility of analyzing and evaluating
differing cultural values - as they show up in the language - and thus
helping them to get rid of their own ethnocentricity, which is in line with
Whorf's claims concerning the impact of culture on language.
to evaluate or judge the differences they perceive but merely to state them
and to reflect upon their repercussions on their own self-image and their
own culture. The foreign language classroom is a perfect environment to
pursue such aims, because learning a foreign language necessarily includes
learning about aspects of a foreign culture. This is also where teaching
methods based on Cognitive Linguistics may come in, as this theoretical
paradigm does not separate language and culture but is based on the insight
that language is always embedded in and part of its cultural environment
and that these two factors are thus inseparably intertwined.
Ifwe take Whorf's insights seriously, we might even want to ask ourselves
how it is feasible at all to teach a foreign language - which is necessarily
rooted in a culture different from that of the learners - without taking the
learners' culture into account. The logical conclusion that imposes itself is
that, first of all, raising the learners' intercultural awareness is a prerequisite
for all successful foreign language teaching, in the sense of making learners
aware of different "isolates of experience", as Lee (1997) would suggest
along with Whorf.3 Thus, for example, even if Germans physiologically
experience much the same natural phenomena when they talk about
"Nebel" and "Dunst" as native English speakers would when they talk about
"haze", "mist", and "fog", the German language has cut up the continuum
at different points than the English language, and German speakers thus do
not talk about exactly the same concepts as English speakers do. The same
applies to the famous example of the multitude of Inuit words for snow:
members of other cultures may well be able to experience the same kinds of
snow but as those are not vital to their everyday lives, their languages have
not lexicalized them. This does not mean that these phenomena do not exist,
just that the language communities chose not to focus on these "isolates of
experience". Second, we have to relativize the dogma of translatability - of
words or categories, of grammatical concepts as well as of elements of cul-
ture - because in translation, words, categories, and concepts, linguistic as
well as otherwise, would simply be transferred to a different conceptual
world where they may be completely out of place (for a Whorfian view on
translation, see also House 2000). Adhering to such principles, foreign lan-
guage teaching should be rethought so as to stress the aspect of the foreign
culture at least equally strongly as that of the lexical and grammatical struc-
tures themselves, and the new trends in foreign language teaching method-
ology as outlined above all seem to point to that direction. Taking Whorf seri-
ously in the foreign language classroom definitely means taking a much more
holistic view on language than the view that is currently present in Germany's
English textbooks and, accordingly, English language classrooms.
Although we should keep in mind, as Lucy (1997: 295) correctly reminds us,
that "linguistic relativity is not the same as cultural relativity", there is still a
very strong interrelation to be found, or as Bickel (2000: 185) formulates it:
"Correlations between linguistic and cultural patterns ... suggest mutual
influence ... , since both speaking and social behavior are publicly shared
activities that are transmitted across generations". And especially in foreign
language teaching, language and culture seem to be very narrowly inter-
twined because the learners' access to the other culture and their means to
discover it are always mediated by the language they are learning, although
one should be aware of the fact that the learners will always be influenced by
their own cultural and linguistic background/s as well. Thus we are dealing
here with a kind of double responsibility, because on the one hand foreign
language teachers are supposed to transmit the language correctly, and on
the other hand they are also supposed to be trying to "translate" and/or to
explain aspects of the target culture without letting the learners' home culture
interfere too much.
How could these claims be satisfied in a foreign language classroom,
apart from reflecting on and possibly acting according to the current trends
in foreign language teaching methodology? One suggestion would be to
turn to some major issues within the framework of Cognitive Linguistics (in
the Langacker-Lakofftradition) and see whether they can help reach at least
some of these aims. As the contributions in this volume show, this seems to
be quite a promising venue. Cognitive Linguistics, with its underlying
humanistic approach, takes a holistic view on language and culture and thus
shares a first common denominator with some of the above-mentioned cur-
rent trends in foreign language teaching methodology.
It is a truism that foreign language learning should not take place in bits
and pieces but aim at integrating all mental capacities of the learner at once
and from the very beginning onwards. Furthermore, a cognitive-linguistic
approach to language teaching and learning also allows for multi-channel
learning insofar as Cognitive Linguistics claims that language is but one of
our mental activities and is directly related to other mental activities such
as, for example, cognition, memory, vision, or senso-motor strategies - the
last also hinting at a connection to the action-oriented approach. Raising the
learners' awareness would then be the major target in this kind of teaching
and learning, and this not only entails raising their language awareness but
also raising their cultural as well as their intercultural awareness so that they
are successfully endowed with the capacity to realize the relativity of con-
cepts within their own culture, as well as those within the culture and the
language they are getting acquainted with. As a consequence, learners may
So far, we have been asking the question of what the notion of linguistic
and cultural relativity could contribute to a strengthening of the learners'
recognition of the learning objective of intercultural competence (as tenta-
tively defined above), or at least to an increased awareness of the fact that
there are indeed major intercultural differences. I will try to answer this
question by drawing on an example from Cognitive Linguistic theory, i.e.,
how can we use the concepts of categorization and prototypicality in order
to achieve the aim outlined above.
First of all, if we consider that Gumperz and Levinson at the very be-
ginning of the introduction to their 1996 volume claim that "culture, through
language, affects the way we think, especially perhaps our classification of
the experienced world" (1996: 1), the association with categorization as it is
discussed in Cognitive Linguistics comes up immediately. The claim is that
different cultures categorize given concepts in different ways - for example,
an avocado is conceptualized as a vegetable in Germany and as a fruit in
other languages, which just goes to show that the boundaries between these
categories are fuzzy - and that such examples are illustrative enough to be
used even with ab initio learners. Of course, the outcome would be even
clearer if we chose cultures farther apart from each other than the German
and the British or American one - but nevertheless, even between these cul-
tures there are many differences to be shown, which furthermore do not
necessarily refer to concrete objects. Thus, for instance, in English we find
the expressions joy and happiness to describe very positive states of mind,
whereas in German we find the expressions gliicklich and zuJrieden (cf.
Wierzbicka in Dirven and Verspoor 1998: 146-148). Although these ex-
pressions do not refer to exactly the same states of mind (as can be proven
by an analysis along the lines ofWierzbicka's Natural Semantic Metalan-
guage approach), they still belong to the same meaning continuum; however,
the two languages in question have cut up this continuum in different ways
and have thus established different categories. Examples of this kind are not
hard to find and are useful to awaken the learners' interest in and openness
for linguistic as well as (inter-) cultural differences.
A central phenomenon within categorization is that of saliency or proto-
typicality. Thus, according to experiments by Rosch and others (cf. Rosch
1978), there is always a "best example" within a category, the so-called pro-
totype. This prototype is extremely culture-dependent - in a culture like
Germany, the prototypical bird would be a sparrow but in a culture like the
USA it is a robin because this bird is much more frequent there than, for
example, a sparrow. Admittedly, this is a rather obvious example; others may
be more subtle. But as prototypes depend on experiences within a culture
and are also often transmitted inside a culture, they have to be seen against
this cultural background.
In this context, research on colors is also of interest. In 1969, Berlin &
Kay already claimed the existence of "focal colors", i.e., best examples or
prototypical colors within a color continuum. These focal colors are culture-
related and also language-related because there are cultures such as the
Papuan Dani culture (see, for example, Heider 1972) - the extreme case that
is often quoted - which have in their language only two color terms and
therefore split up the color continuum in a different way than, for example,
the English language or the German language do, which possess eleven
focal color terms each. Thus, the linguistic color continuum is not a univer-
sally given fact but - as Lucy (1997) also claims - depends on the cultural
background and influences language and culture at the same time. For in-
stance, in languages with fewer basic color terms, people's color memory
seems to be less accurate than in languages with more basic color terms
(Lucy & Shweder 1979, 1988). As the focal color terms are usually mono-
morphemic and are among the first lexemes to be learnt when learning a
first or foreign language, the differences between two cultures can be pointed
out by focussing on the different referents for those terms. However, at first
glance, this does not seem to work when focus sing on the English vs. the
German language as they seem to have the same focal colors.5 However, if
we take color expressions such as "blue-eyed boy" or "red tape" into ac-
count (see Niemeier 1998) as well as the non-existence of possible German
counterparts, we are able to explain the culture-boundedness and thus the
non-translatability of these expressions.
The question remains of what the issues of categorization and prototypi-
cality have to offer for second or foreign language learning. First of all, let us
stress again that their specific strength is to help raise language awareness.
Monolingual people tend to think - in a rather ethnocentric way - that the
terms and prototypes found in their mother tongue describe universal issues,
and only via the detour of learning a foreign language may learners become
aware of how differently languages can be structured 6 and how different the
cultural concepts behind those linguistic structures are. The above mentioned
and similar examples may all be used even in beginners' classes and it is
decisive to use them exactly there, in order to show the learners from the very
beginning that they should not expect one-to-one meaning relationships or
structural similarities between their mother tongue and the foreign language.
This is an important issue as well when it comes to translation, because such
cultural differences in categorization and prototypicality as named above
implicitly negate the existence of a one-to-one meaning transfer, not to
mention the different associations with cultural norms and values that most
lexical items entail. This is a crucial issue for the learners to internalize be-
cause it will help them deal with more intricate concepts later on, and because
it allows them to construct their knowledge of the foreign language and the
foreign culture more independently from their own culture and language.
Thus, learning to realize and even to make sense of the cultural concepts
behind the words of the foreign language may also make learners more con-
scious of their own language, and therefore this way of learning (about)
intercultural differences often induces a process of reflecting on one's own
culture and cultural image for the first time. Intercultural understanding
does not necessarily mean stating what is different between the cultures, but
it focuses on showing that there ARE such differences, that one should be
aware of them, and that one should use the foreign language accordingly,
i.e., not referring to one's own cultural background but being aware of the
foreign cultural background. In a nutshell, learners should especially learn
to see their own language and culture as only one among many sensible
experiences of the environmental world. As a very trivial and often quoted
example, we can take the famous "How do you do?" utterance, which does
not call for a detailed report on one's health problems, but as an act of
Two other, interrelated areas within Cognitive Linguistics that can be made
fruitful for foreign language teaching - especially on a more advanced level
- are conceptual metaphors and metonymies. Metaphors as well as
metonymies are so heavily entrenched in the foreign culture that they can
not be properly interpreted or explained without referring to their cultural
background. Therefore they seem to be a good means to induce knowledge
about the culture in question or at least to have the learners baffied by their
non-translatability and thus raise their interest. As to metaphors, Cognitive
Linguistics is not primarily interested in innovative, "creative metaphors"
(see also Lakoff and Turner 1989) as they are dealt with in traditional litera-
ture studies and that exploit the same conceptual metaphors as folk
metaphors do, but focuses on the strategies that underlie both these creative
metaphors and the so-called conventional metaphors, which are not used
consciously as metaphors anymore. These fossilized metaphors are never-
theless very efficient, as they may even be able to influence our thought
processes. Not only do we use them subconsciously but we also think in
terms of these metaphors without being aware of it, as they are so much a
part of our language that we do not pay any special attention to them; but
there is indeed a structuring behind the widespread usage of linguistic or
everyday metaphors.
Lakoff coined the term "conceptual metaphor" to highlight the fact that
within our conceptual systems we have a metaphorical structure which allows
or even constrains us to follow certain patterns of thought. A frequently
quoted example of a conceptual metaphor is TIME IS MONEY and the meaning
transfer (or mapping) from "time" to "money" is not a conscious one but a
cognitive predisposition within our culture. The concept of TIME IS MONEY
comes to the fore in many linguistic instantiations, such as I spent an hour
with my grandmother yesterday, This flat tire cost me an hour, I lost a lot of
time in the traffic jam and many others, as there is an unlimited number of
expressions that we can use. However, the mental structuring behind all of
these expressions is the TIME IS MONEY concept in our minds. We talk
about time in money terms because time is such an abstract concept that it is
difficult to express directly, so we turn towards a more concrete domain of
knowledge - such as money - and map it onto time.
From a Cognitive Linguistic perspective, metaphors function by mapping
the meaning structure of a more concrete source domain onto the as yet
dimly seen conceptual structure of a more abstract target domain in order to
facilitate understanding of this second domain in some way. Thus, we see a
similarity between two different domains (in our example, money being the
concrete source domain and time being the abstract target domain) and map
characteristics from the source domain onto the target domain. Generally this
does not happen in a conscious way, it is part of our linguistic and cultural
socialization. In a culture where people are not paid by the hour, the month,
or the year, such a mapping would presumably not come into existence.
However, in this respect, there is no big cultural difference between Anglo-
American cultures and the German one, so we find this conceptual meta-
phor in German as well, although the linguistic instantiations are not exactly
the same. Thus we do not "spend" time in German, but we can "Zeit inves-
tieren", "Zeit gewinnen", "Zeit verlieren", "Zeit sparen", "Zeit verschwen-
den" and "ein biBchen Geduld kann sich auch auszahlen".
The TIME IS MONEY metaphor is, of course, not the only way to talk
about time in English, as there are other conceptual metaphors connected to
the target domain of time, depending on what we want to express. We may
also use the conceptual metaphors TIME IS A MOVING OBJECT (in examples
such as The years flew by, or It is getting close to bedtime), or TIME IS A
THIEF (in examples such as The passing of the years has robbed me of my
beauty, or when Milton in Sonnet 7 calls time the subtle thief ofyouth), or
TIME IS A CHANGER (in examples such as Over the last years, I have
become a lot more patient, or Time heals all wounds) or as many other
things (cf. Lakoff/Turner 1989). This quantity of conceptual metaphors is
due, on the one hand, to the fact that time is such an important and salient
concept for us that we need various ways to talk about it and, on the other
hand, to the fact that no culture is homogeneous and we need a choice of
how we want to conceive of time, depending on our social, religious, or
individual backgrounds. However, this choice is not an unlimited one be-
cause we are not free to choose our conceptual metaphors as they are part of
our linguistic and cultural heritage, and we can only choose among those
that exist within the target culture.
If we take the example of "life", a concept that is hard to define but
which is nevertheless very central for us humans, we find many different
conceptualizations. One of Lakoff's best-known examples is LIFE IS A
JOURNEY (Lakoff1980). A journey is a very concrete event which all of us
have some experience with, whereas the more abstract concept of "life" is
hard to describe, apart from a purely biological definition. But once we try
to grasp its meaning by mapping our everyday knowledge of "journeys"
onto the concept of life, it becomes much less obtuse and we may see life as
having a beginning, as pursuing a certain route, as possibly involving side-
tracks and obstacles, and as having a final aim or an end. In this way, we are
able to talk about "life" and make ourselves understood, at least inside the
culture. Even if we only rarely think of life explicitly in terms of a journey,
this conceptual structuring nevertheless exists in our minds, since it comes
to the surface in a number of concrete metaphors.
Even the Bible makes abundant use of this conceptual metaphor, for
instance when it talks about the "good way" and the "evil way" in the Old
Testament. However, different belief systems have chosen to see "life" from
other perspectives, such as the Hindu culture where life is seen as a transi-
tory stage to a better existence (also present in Christian culture in the belief
in a better "life" after death). Lakoff and Turner (1989) quote 17 different
basic metaphors for "life" that are present in the Western culture, which,
according to them, is an astonishingly small number for such an important
concept. The different conceptualizations of "life" mainly rely on the every-
day experiences made in a certain cultural context, which are handed down
from generation to generation. Also, new creative metaphors ultimately rely
on these conceptual metaphors because they make use of the schemata
underlying the conceptual metaphors and extend them in innovative ways
(cf. Lakoffand Turner 1989).
How can this approach be made fruitful for the foreign language class-
room? From a functional point of view, we can state that metaphors are
often used for manipUlative purposes, be it in advertising or in political
propaganda of all kinds. If news coverage on a war or on political unrest
metaphorically frames one party as "the bad guys" and the other one as "the
good guys" and does so by exploiting common conceptual metaphors, we
are normally not even aware of this kind of manipulation, so that we do not
have many possibilities of seeing the events in any other light. There are
examples of this concerning the news coverage on the Gulf War or also the
September 11 event. The manipulative potential of metaphors, then, is one
of the reasons why learners should at least be made aware of the notion of
conceptual metaphors, so that they may have a chance to see as well as react
to the manipulative potential of language, and that they may realize what a
powerful and flexible tool language really is. This is, of course, not only
true for a foreign language but also for the native language, since, via the
alienating effect of realizing how the foreign language works, one may also
arrive at a more realistic view of the native language. Working with
metaphors in the foreign language classroom therefore helps raise the learn-
ers' awareness for structuring principles within language and thought and
for the cultural differences within this structuring.
Metaphors always involve cultural knowledge, which is another reason
to focus on metaphors in the foreign language classroom because they can
raise the learners' intercultural awareness and may be used for a contrastive
approach. We can see this, for example, when learners actively work with
metaphors in comparing conceptual metaphors, as well as their linguistic in-
stantiations, in English and in German texts of any kind and of their own
choice - because metaphors are everywhere! - and compare and discuss
their results. It may also be connected to the analysis of literature or used as
a conscious strategy in creative writing. Working with metaphors also
entails an affective element, as such an approach may raise the learners'
motivation because it is fun to uncover manipulation and may induce learn-
ers to look more closely at language and to decipher "hidden meanings". In
this way, they can also monitor their own language output in a more
meaningful way. The main point to stress is that working with metaphors is
a way to understand how language works, how it changes and how meanings
are created and extended. In this way, learners may develop a new perspec-
tive on language in general.
Conceptual metonymies are even more pervasive than conceptual meta-
phors and, as with metaphors, more often than not people are not even aware
of using them. Metonymies come in a great variety of shapes, each one of
them highlighting another relation between the referent as such and the
metonymy denoting it. In each case, the most salient aspect of the context is
verbalized to evoke the whole concept or scene. To give an example, in the
WHOLE FOR PART metonymy as in "The Rolls-Royce left the gas station
without paying", where it is of course the less salient driver of the more
salient Rolls-Royce who did not pay, we immediately recognize the whole
scenario of events.
As with metaphors, the coding and decoding processes connected to
metonymy are largely determined by extralinguistic experiences and conno-
tations. Therefore, intercultural differences will show up in the existence or
the "congealing" of such metonymies, because different aspects are seen as
the most salient ones, and thus different metonymies are conceptualized.
These different aspects have much to do with the cultural background in
which the language in question is spoken. Thus, when we find expressions
with "green" in English (or closely related cultures) which refer to ecologi-
cal preservation (green party, Greenpeace) via the meaning link of "green"
as a metonymy for nature (green belt, green thumb). This is due to the pre-
dominance of naturally occurring green entities like leaves, plants, or grass
(cf. also Niemeier 1998). However, such metonymies would probably not
develop in cultures where there is not much green in the daily environment.
Making the learners aware of such differences might be a useful inroad to
raising their intercultural awareness. Another reason why the concept of
4. Conclusion
It cannot be stressed enough that the methods outlined in the text are
designed - and in my opinion well able - to help learners realize the close
connection between a language and its culture, and to see how the meaning
networks in the foreign language are constructed and how a language adds
new concepts and meanings almost every day. This way of looking at the
foreign language does not just entice the learners to learn concepts by heart,
but it offers them an insight into the ways the language works. Especially
metaphors and metonymies are packed with cultural knowledge. Their
meanings are not ad hoc meanings at all, but they are all motivated by dif-
ferent aspects of the culture in question, and thus able to transfer information
about this culture.
Therefore, using some of the insights and principles of the cognitive
linguistic approach in the foreign language classroom or even - for a start -
inserting cognitive linguistic elements into the "normal" teaching methods
seems to be particularly relevant for vocabulary training and for intercul-
turallearning for advanced learners. Furthermore, the examples mentioned
above are also relevant for grammar teaching, because in grammar we also
find categorization and prototypicality and even metaphors, such as, for
example, metaphorized prepositions (see Niemeier 2001). Furthermore, as I
have tried to show, the methods discussed are compatible with most of the
current didactic theories such as the concern with awareness raising, the
learning objective of intercultural competence, the learner-centered approach
(Le., autonomous learning), multi-channel learning, the tendency towards a
more holistic kind of teaching and learning, the action-oriented approach to
learning and teaching, as well as the focus-on-form approach. What all of
these approaches have in common is that they do not compartmentalize dif-
ferent aspects of the language (as, for example, grammar teaching as
opposed to or at least as more or less unrelated to vocabulary training) but
that they are based upon a concept of language that regards language in a
holistic way as inseparably connected with culture and with people's mental
as well as bodily experiences.
Such an approach to teaching and learning foreign languages seems to be
in line with Whorf's ideas insofar as it helps learners to not try to look for
equations between their native language and the foreign language, but that
they are open for other conceptualizations within the foreign language and
its culture. As pointed out above, a metonymy may arise when a salient
detail in a given domain is taken to represent the whole scenario or domain.
However, the saliency of an object, event, or relation is construed by the
language user and is not objectively or automatically given. The language
user, on the other hand, does not live in a culture-free environment and is
thus influenced by the surrounding culture, if only subconsciously. There-
fore, the meaning extensions are bound to mirror the underlying culture to a
certain degree. For the foreign language learner this means that these under-
lying and possibly contrasting factors - closely bound to the language -
have to be brought to mind, and that possible discrepancies between the
learners' home culture and the foreign target culture have to be dealt with
much more consciously.
Acknowledgment
Notes
1. This text can only account for the German point of view, as educational systems
vary markedly in international comparison.
2. Currently a generally accepted defmition of "intercultural competence" does not
seem to be available, therefore I will not try to give one here (for further discus-
sion, see, for example, Byram 1997, House 1997, House and Edmondson 1998).
My own working definition of "intercultural competence" is an understanding
of the major general differences between the target culture and the home culture,
and thus a prevention of prejudices and a neutral re-evaluation of both cultures.
This working definition is relatively close to Kramsch's "third domain" (see
Kramsch 1993), which, however, seems to me too static a concept.
3. Lee (1997) defines these "isolates of experience" as following: "An isolate of
experience can be figure or ground. An isolate is any phenomenon which we
can abstract away from the matrix of its occurrence as something separately
perceivable or noticeable. Isolates can occur in any sensory domain."
4. The 6th ICLC conference in Stockholm, July 1999, featured three plenaries on
language acquisition and a one-day workshop on "Cognitive linguistics and lan-
guage acquisition" organized by AchardlNiemeierlSinha. At the 7th ICLC con-
ference in Santa Barbara, July 2001, several papers on the topic in question were
presented. Furthermore, the 28th LAUD Symposium (organized by Piitzl
Niemeier) in Landau/Germany, March 2000, was completely devoted to the topic
of "Cognitive Linguistics: Second Language Acquisition, Language Pedagogy,
and Linguistic Theory", and GURT 2003 (Georgetown University, Washington)
has as its general topic "Cognitive and Discourse Perspectives on Language and
Language Learning".
5. However, this is by no means clearly determined. Currently there is research in
progress which tries to fmd out the subtle meaning differences between focal
color terms in a couple of European languages, among them German and English
(personal communication by Robert MacLaury).
6. Again, although this may work even better with languages that are culturally
further apart from each other than English and German are, between these two
languages we can also find salient structural differences, which influences the
language users' conceptualization processes and linguistic output, such as, for
example, the progressive aspect in English verbs, which is not directly marked
on German verbs - or, to use cognitive linguistic terminology, the notion of
boundedness vs. unboundedness in verbs (see Niemeier 2001 and forthcoming).
References
House, Juliane
1997 "A contrastive approach to developing intercultural competence".
Sprak og Marked, 15,3-10.
2000 "Linguistic relativity and translation". In: Martin Piitz and Marjolijn
Verspoor (eds.), 69-88.
House, Juliane and Willis Edmondson
1998 "Interkulturelles Lemen - ein iiberfliissiger Begriff". Zeitschrijt for
Fremdsprachenforschung 9: 2, 161-188.
James, Carl and Peter Garrett (eds.)
1995 Language Awareness in the Classroom. London: Longman.
Kramsch, Claire
1993 Context and Culture in Language Teaching. Oxford: Oxford Univer-
sity Press.
Lakoff, George and Mark Johnson
1980 Metaphors We Live By. Chicago: University of Chicago Press.
Lakoff, George and Mark Turner
1989 More than Cool Reason: A Field Guide to Poetic Metaphor. Chicago:
University of Chicago Press.
Lamb, Sydney M.
2000 "Neuro-cognitive structures in the interplay of language and thought".
In: Martin Piitz and Marjolijn Verspoor (eds.), 173-196
Langacker, Ronald
1991 Concept, Image and Symbol. The Cognitive Basis of Grammar.
Berlin: Mouton de Gruyter.
Lee,Penny
1997 "Language in thinking and learning: Pedagogy and the New Whorfian
framework". Harvard Educational Review 67: 3,430-471.
van Lier, Leo
1996 Interaction in the Language Curriculum. Awareness, Autonomy and
Authenticity. London: Longman.
Little, David
1994 "Learner autonomy: A theoretical construct and its practical appli-
cation". Die Neueren Sprachen 93, 430-442.
L6ffier, Renate
1996 "Ganzheitliches Lemen: Grundlagen undArbeitsformen". In Gerhard
Bach and Johannes-Peter Timm (eds.), 42-68.
Lucy, John A.
1992 Language Diversity and Thought. A Reformulation of the Linguistic
Relativity Hypothesis. Cambridge: Cambridge University Press.
1997 "Linguistic relativity". Annual Review ofAnthropology 26,291-312.
Lucy, John A. and R.A. Shweder
1979 "Whorf and his critics: Linguistic and noniinguistic influences on
color memory". American Anthropologist 90, 923-931.
1988 "The effect of incidental conversation of memory for focal colors".
American Anthropologist 33, 137-146.
McCarthy, Michael
1991 Discourse Analysis jor Language Teachers. Cambridge, UK:
Cambridge University Press.
Niemeier, Susanne
1998 "Colourless green ideas metonymise furiously". In: Friedrich Ungerer
(ed.), Kognitive Lexikologie und Syntax. Rostocker Beitriige zur
Sprachwissenschaft 5, Rostock: University ofRostock, 119-146.
2001 "Zur Beziehung zwischen Lexik und Grammatik". In: Dagmar
Abendroth-Timmer and Gerhard Bach (eds.). Mehrsprachiges
Europa. Tiibingen: Narr, 305-326.
Forthcoming
"BoundednesslUnboundedness - Blick durchs Schliisselloch. Ange-
wandte Kognitive Linguistik fur den Englischunterricht".
Niemeier, Susanne and Rene Dirven (eds.)
2000 Evidencefor Linguistic Relativity. Amsterdam: Benjamins.
Palmer, Gary.
1996 Toward a Theory of Cultural Linguistics. Austin: University of Texas
Press.
Panther, Klaus-Uwe and Giinter Radden (eds.)
1999 Metonymy in Thought and Language. Amsterdam: Benjamins.
Pica, Teresa
2000 "Tradition and transition in English language teaching methodology".
System 28: 1-18.
Piitz, Martin and Marjolijn Verspoor (eds.)
2000 Explorations in Linguistic Relativity. Amsterdam: Benjamins.
Rosch, Eleanor
1975 "Cognitive representations of semantic categories". Journal ofEx-
perimental Psychology: General 104, 192-233.
1978 "Principles of categorization". In Eleanor Rosch and Barbara B.
Lloyd (eds.), Cognition and Categorization. Hillsdale: Lawrence
Erlbaum, 27-38.
Selinker, Larry
1992 Rediscovering Interlanguage. London: Longman.
Timm, Johannes-Peter (ed.)
1995 Ganzheitlicher Fremdsprachenunterricht. Weinheim: Deutscher Stu-
dien Verlag.
Wendt, Michael
1996 Konstruktivistische Fremdsprachendidaktik. Lerner- und handlungs-
orientierter Fremdsprachenunterricht aus neuer Sicht. Tiibingen: Narr.
Whorf, Benjamin L.
1956 "The relation of habitual thought and behavior to language". In John
(1941) B. Carroll (ed.), Language, Thought, and Reality: Selected Writings
by Benjamin Lee Whoif. Cambridge, Mass.: MIT Press, 134-159.
Wolff, Dieter
1994 Lernerautonomie. Themenheft: Die Neueren Sprachen 93(5).
Peter Grundy
1. Introduction
This chapter reviews the ways in which the figure / ground gestalt has been
applied in linguistics and considers its relevance to second language teaching
and learning. In particular, it is argued that the figure / ground gestalt under-
mines the discrete-item-syllabus assumption on which instructed second
language acquisition is typically based and that it is precisely its figure /
ground properties, especially as revealed in its metapragmatic dimension,
that make language comprehensible, and thus learnable.
The commonsensica1 view that language, like any other "subject", is teach-
able may have some credibility in pedagogy, but has been largely rejected
by both generative and non-generative applied linguists who have come to
argue over the last ten or fifteen years that language is a learnable behaviour
rather than a teachable subject. Writing from a generative perspective,
Wong-Fillmore puts the position clearly: "In current theories of language
'Iearnability', a principle of 'teach-ability' is unnecessary" (1989: 311).
For generativists such as Wong-Fillmore, this is a natural fallout of the
innatist view that language acquisition is biologically determined and thus a
system-internal phenomenon. This position is supported by the good results
obtained by generativists working with learner data which demonstrate that
learners appear to know more than they have been taught (Bley-Vroman,
Felix and loup 1988).
For non-generative, pedagogically oriented applied linguists, the privileg-
ing of learner discovery over teacher instruction and of learner self-expression
over the memorization of model structures has also promoted the learner as
the significant determinant of acquisition. As Crandall writes, "Since the
mid 80s, discussions of effective language instruction have shifted from an
emphasis on teacher-centred to an emphasis on learner-centred classrooms
and from transmission-oriented to participatory or constructivist knowledge
development" (1999: 226).
Rejecting teachability in favour of learnability on the grounds that lan-
guage acquisition is system-internally determined, a reflex of a language
dedicated module in the mind, enables us to speculate on the extent to which
universal grammar is available to post-puberty language learners and on the
influence of prior knowledge on second language acquisition. But substituting
a learnability for a teachability hypothesis does not tell us how the acquisition
process can be assisted in the classroom, beyond vague claims about the
need for comprehensible input and a low affective filter setting (Krashen
1985). Yet we need to know not just that language acquisition processes are
determined by cognitive structures and likely to produce output data accord-
ingly, but, more to the point, how the language input data needs to be pre-
sented to these cognitive structures for successful acquisition to occur.
If the structure of the brain is uniquely adapted to understanding the
nature of the phenomena it encounters in the world, including linguistic phe-
nomena, there is a sense in which it is not just the learner, but rather language
itself, which teaches the learner language. This view obliges us to look at
language not as a counterpart to innate biological structures, but as one of a
The insight that figures are associated with discreteness, shape, and singular-
ity, whereas diffuseness is the principal characteristic associated with ground,
is owed to the pioneering experimental work of Rubin (1915) in the field of
visual perception. What we "see" is the figure. We assume that the contour
marking the boundary of figure and ground belongs to the figure and not to
the ground against which the figure is, consequently, profiled. Figures give
an impression of solidity, closeness, and density of colour relative to ground.
The ground appears to continue uninterrupted behind the figure. It is the
figure and not the ground which is remembered.
In relation to the "actual" world, these perceptions are, strictly speaking,
illusions. They are, therefore, direct reflections of cognitive processes which
impose, even on a flat surface such as the page you are now reading, the
impression that some parts of what you see, in this case the graphemic sym-
bols, are closer to you than the background (I cannot avoid the metaphor) on
which they are printed. The salient shapes to which your attention is drawn
are those of the symbols and not of the background page. Moreover, you can
easily reproduce the symbols, but would have immense difficulty reproducing
the shape which shares a common contour with them.
The issue then is whether linguistic phenomena too have salience in a
figure / ground manner and whether the figure / ground gestalt has implica-
tions for language learning and teaching.
figure / ground gestalt, Mukarovsky's work and that of his colleagues never-
theless clearly presupposes such a gestalt. For Mukarovsky, poetic language
consists of the fore grounding of phonological, syntactic, or even semantic
features, whose resulting prominence tends to push meaning into the back-
ground. Thus the formal means of expression rather than what is expressed
is the salient figure. What we notice in the case of poetic writing is not what
is conveyed but how it is conveyed.
Working in a different tradition and in a sense anticipating the work of
cognitive linguists, Wall ace (1982) attempts to draw together and expand on
the considerable body of existing descriptive work on linguistic categories
which, he argues, demonstrate the figure / ground polarity. He observes that
languages have a range of contrasting grammatical forms (perfective / im-
perfective, eventive / non-eventive, definite / indefinite, etc.) which are pre-
dominantly oriented to expressing figure / ground relations. Thus, Wallace's
use of figure / ground is oriented to the explanation of discoursal phe-
nomena: a speaker has the means at each point in a discourse to foreground
some element of propositional meaning in relation to some other element.
The most influential application of the figure / ground gestalt in linguis-
tics, however, is found in Langacker's work where it underlies the profile /
base relation, one of the fundamental principles on which the cognitive,
usage based account oflanguage rests:
A predication always has a certain scope, and within that scope it selects a
particular substructure for designation. To suggest the special prominence of
the designated element, I refer to the scope of the predication and its desig-
natum as base and profile [Langacker's emphasis] respectively. Perceived
intuitively, the profile (in the words of Susan Lindner) "stands out in bas-
relief" against the base. The semantic value of an expression resides in neither
the base nor the profile alone, but only in their combination; it derives from
the designation of a specific entity identified and characterized by its position
within a larger configuration (Langacker, 1987: 183).
the figure / ground properties of pragmatic phenomena have often been noted.
In Talmy's account of complex sentences, for example, figure is seen as a
variable (1978: 628ft) in relation to a ground provided in the subordinate
clause. Thus clauses introduced by after and before, and phrases introduced
by during provide the ground, or presupposition, in relation to which the rest
of the sentence is seen as a figure. This assertion-as-figure / presupposition-
as-ground position is also taken up by Levinson (1983: 180).
Even more strikingly perhaps, at the level of the word, Hanks (1992) argues
that deictics among pragmatic phenomena uniquely capture the relation offig-
ure to ground in a single linguistic expression. Thus what a demonstrative
points to as a figure in an expression like you or this year is related to the
indexical ground or deictic anchoring point of the speaker (and sometimes
of the hearer) at the time when and in the place where the utterance occurs.
Each of these accounts, whether seeing ground as co-text (Langacker,
Talmy) or context (Hanks) or as a linguistic option which is not chosen
(Mukarovsky, Wallace), demonstrates the linguistic reflexes of a fundamen-
tal processing strategy in which salience is perceived as a relation of figure
to ground. And in so doing, the accounts generalize a perceptual theory to
the understanding of language. However, what they do not do is explain
how the grounding process itself occurs. It is to this problem that we turn in
the following section.
A semantic function that constitutes the final step in the formation of a nomi-
nal or a finite clause. With respect to the fundamental "epistemic" notions
(e.g., definiteness for nominals, tense/modality for clauses), it establishes the
location vis-a-vis the ground of the thing or process instance designated by
the nominal or the clause overall (1991 a: 549).
vant (i.e., what might be meant by uttering it) when recovery of the set of
grounding propositions (aI, a2, a3... a n have private lives), is not triggered by
the procedural "even". Thus procedural encodings provide the hearer with an
indication of how to provide contexts in relation to which the implicatures
conveyed by conceptual meanings may be inferred. Put simply, they enable
a hearer to recover a context or ground in relation to which a sentence or an
utterance can be a figure.
The discussion in the previous paragraphs of the way grounding is handled
in Relevance Theory provides a particularly clear account of the role of just
one category of metapragmatic grounding phenomena, discourse particles.
In a survey paper, Verschueren (2000) lists many other metapragmatic means
of grounding, including metapragmatic descriptions in the form of speech
act and performative verbs, self-referential expressions, sentence adverbs,
hedges, deictics, evidentials, aspect, modality, prosody, code switching, and
implicit voices. Verschueren argues that metapragmatic phenomena should
not be regarded merely as linguistic objects (as the list above might suggest)
but rather as a dimension of language, so that the way any proposition is
conveyed provides evidence of the metapragmatic awareness of the speaker
and of the manner in which the proposition is grounded in the context in
which it occurs.
what is denoted, we need not only a demonstratum but also an origo. Without
presupposition, whether semantic or pragmatic, assertions and other speech
act types are just that, types and not grounded instances, and thus not readily
interpretable. Perfectives need an imperfective discourse for their effect.
Speech acts require their containing speech events. And conceptual mean-
ings frequently require the grounding procedural encodings which enable
the recovery of the base in relation to which their relevance is understood.
There is a sense in which it might be said that without ground, there is no
figure. While this does not seem a very controversial claim to a cognitive
linguist, it seems to not have been widely accepted in language teaching
methodology.
All too often language teaching materials consist of "bare" instances of
language types or "bare", ungrounded dialogues. Although Text A below is
not, as will be seen, an authentic coursebook dialogue, it will serve as
broadly representative of the type. You need to know that the first speaker,
D, is the warden of a student hall of residence and that the anaphor "it" in
his first turn refers to the theft of some butter, while the second speaker, N,
is a student and that the same anaphor in her first turn refers to the butter
that was stolen:
Text A
Text B
D how's it doing
S yeah no erm Nicole wants to no we need a word with you
D oh Jight_
N _no _ someone's been nicking stuff out the fridge (2.0) so
D which one
or what she chooses to leave implicit" (1995: 281). Thus a question such as
Are you working this afternoon? might be understood as a request for a lift,
as conveying that the speaker did not expect the addressee to be working, or
indeed as virtually anything else that the utterance Are you working this
afternoon? taken together with the non-linguistic context or ground might
seem to support (Grundy 2000: 72). Thus the utterance itself is part of the
ground. Yet teachers are frequently anxious to avoid implicature, preferring
instead to deal only in entailment.
In a pedagogic situation, the contextual resources or ground are supplied
by an addressee who is also a language learner. The teacher who is able to
accurately estimate the contextual resources available will therefore be able
to test the learner's understanding of what is said if the learner demonstrates
this understanding by drawing the appropriate inferences, as demonstrated
by an ability to sustain the exchange.
The reality, however, is that speakers often use procedural encodings
alongside conceptual encodings to indicate to the addressee how a profiled
utterance is to be related to a previous part of the discourse, which is there-
fore to be treated as ground. Or where the available range of contexts might
be very great, a procedural encoding can be used to constrain the likely con-
texts that might be considered, which once again assists the addressee in
determining the ground in relation to which an utterance is to be understood.
Procedural encoding may also function to indicate how an utterance is to be
treated as a ground for the following utterance or turn - the Japanese sentence
final particles ne, yo, and yone are notable examples of procedurals which
function in this way. In recognizing the important function of procedural
encodings as grounding devices, we now see that they enable the right inter-
pretation of an utterance by pointing the addressee to the ground required
for understanding it. Thus, as we saw when we considered the talk exchange
in the previous section, exchanges without regulatory metalanguage such as
Text A are much more difficult to interpret than exchanges with a rich
accompaniment of metalinguistic cues. This is difficult for materials devel-
opers to accept because, in their natural focus on what is to be learnt, they
tend to see only figure and to overlook that without ground the figure is not
readily interpretable.
In the final subsection of this part of the paper I will argue that as simulta-
neous encodings of figure and ground, indexicals should be prominent in a
It is widely held that second or foreign language learners should start with
what is "easy" and then proceed to what is "difficult". Defining "easy" and
"difficult" is not unproblematic, however.
Most teachers of English would appeal to morphological criteria in rating
the present simple "easier" than the present perfect or the present continuous
(or even perhaps than the past). Similarly, it seems perverse to argue that the
comparative and superlative forms of adjectives are "easier" than the base
form. However, this position has been complicated by work on accessibility
hierarchies and linguistic universals which suggests that the existence of a
marked form implies the existence of an unmarked form within the same
hierarchy. The effects of teaching marked forms such as less-accessible rel-
atives before unmarked has been the subject of some experimentation
(Doughty, Eckman et al.). The results of these experiments do indeed sug-
gest that learners infer the existence of less-marked forms and that teaching
marked forms first may therefore be an economical way of helping learners
to acquire a language.
The figure / ground gestalt, at least as Wall ace interprets it, adds a new
dimension to the easy / difficult issue by suggesting that some linguistic cat-
egories are more salient than others. Should we therefore teach salient cate-
gories early and leave less salient ones until later? Thus we might want to
teach (or increase the frequency of the occurrence of) count rather than
mass nouns, definite rather than non-definite descriptions, first and second
rather than third persons, perfective rather than imperfective aspects, etc.
(Wallace 1982: 212). Or if it could be shown that there is in fact a salience
"hierarchy", then we might argue that less salient members of the hierarchy
should be taught early on the grounds that they imply the existence of more
salient members in the same hierarchy.
Although to base instructional practice narrowly on a salience hierarchy
hypothesis of this sort is to take a rather instrumental view of research find-
ings, we need to acknowledge that "take-away" understanding of this kind
is welcomed by busy teachers who necessarily rely on applied linguists to
provide the rationale for classroom practice.
In our earlier discussion of the figure / ground gestalt, we noted that it is the
figure and not the ground which is remembered. If we extrapolate this to
language and in particular to Talmy's distinction of main sentence assertion
as figure and embedded sentence presupposition as ground, we have an
interesting question to resolve: which are we to practise, the salient item or
the less salient? (In asking this question, I make the perhaps insufficiently
questioned assumption that practice is a vital part of language learning.)
To take a concrete case, in audiolingual-type drills, the teacher has to
decide whether to vary the figure or the ground. Imagine a teacher prompt-
ing controlled practice around the structure I + <verb> + <adverb> +
<preposition phrase> and working away from the textbook model I work
hard at school, where I work hard is seen as figure (trajector) and at school
as ground (landmark).
The first question is whether this structure should be used to practise the
figure I work hard or the ground at school. And even if we knew how to
resolve this issue, there is a further question. Imagine the teacher decides to
cue a drill with flashcards containing words or pictures which indicate the
part of the sentence to be substituted. Should it be the figure which is
replaced (1 work hard at school->l write fast at school->l sit quietly at
school-> You work hard at school, etc.) or the ground (1 work hard at
school->l work hard at home->l work hard in the gym, etc.)? We may have
varying opinions about this simple example, but it raises a fundamental
question - what is the relationship between varied repetition, figure and
ground, and learning?
7. Conclusion
Acknowledgement
References
Blakemore, Diane
1987 Semantic Constraints on Relevance. Oxford: Blackwell
Bley-Vroman, Robert, Stephan Felix, and Georgette Ioup
1988 The accessibility of universal grammar in adult second language
learning. Second Language Research 4,1-32
Chafe, Wallace L
1994 Discourse, Consciousness and Time: The Flow and Displacement of
Conscious Experience in Speaking and Writing. Chicago: University
of Chicago Press.
Crandall, JoAnn
1999 Cooperative language learning and affective factors. In Jane Arnold
(ed.) Affect in Language Learning Cambridge: Cambridge University
Press.
Doughty, Catherine
1990 Second language instruction does make a difference: Evidence from
an empirical study of SL relativization. Studies in Second Language
Acquisition 13.4, 431-69.
Eckman, F.R., L. Bell, & D. Nelson
1988 On the generalization of relative clause instruction in the acquisition
of English as a second language. Applied Linguistics 9.1.
Grundy, Peter
2000 Doing Pragmatics (2nd ed.). London: Arnold.
Halliday, Michael A. K.
1976 System and Function in Language. Oxford: Oxford University Press.
Hanks, William H.
1992 The indexical ground of deictic reference. In Duranti Alessandro &
Charles Goodwin (eds.) Rethinking Context. Cambridge: Cambridge
University Press, 43-76.
Howatt, A.P.R.
1984 A History of English Language Teaching. Oxford: Oxford University
Press.
Hymes, Dell H.
1971 On communicative competence. In J. B. Pride & Janet Holmes (eds.)
Sociolinguistics. Harmondsworth: Penguin 1972, 269-93.
Jakobson, Roman
1960 Linguistics and poetics: Closing statement. In Thomas A. Sebeok
(ed.) Style in Language. Cambridge, Mass.: MIT Press, 350-77.
Krashen, Stephen D.
1985 The Input Hypothesis. London: Longman.
Langacker, Ronald W.
1987 Foundations of Cognitive Grammar: Volume 1, Theoretical Pre-
requisites. Stanford, Calif.: Stanford University Press.
Cliff Goddard
This paper is concerned with the potential applications of one stream of cog-
nitive linguistics - the cross-linguistic semantics of Anna Wierzbicka and
colleagues - to the teaching of ethnopragmatics (cultural pragmatics). To
begin with, therefore, some remarks on conventional approaches to linguistic
pragmatics are in order.
Pragmatic differences between languages have been addressed within
several frameworks, including the ethnography of communication, con-
trastive pragmatics, and linguistic anthropology. In general, work within
these different frameworks assumes that in any speech community there
will be certain shared understandings (variously termed "rules of speaking",
"norms of interaction", "preferred discourse strategies", etc.) about how it is
appropriate to speak in particular, culturally construed situations. The first
task in ethnopragmatic description is, therefore, to identify and describe
culturally preferred speech patterns. In most work there is a further step as
well, for it is widely recognised that language-specific speech patterns are
frequently, if not invariably, linked with the cultural values and attitudes of
the people concerned. The second task is, therefore, to spell out these links.
Both tasks are hindered, however, by an often unrecognised problem,
namely, weaknesses in the descriptive metalanguage which is being
employed (Wierzbicka 1991).
To describe culturally preferred styles of speaking, it is standard practice
to employ an inventory of terms such as "directness", "formality", "involve-
ment", and so on. Though such terms can be valuable and useful up to a point,
they are somewhat vague and tend to be used with rather different meanings
by different authors. For example, when Japanese speech patterns are con-
trasted with English ones, the Japanese are described as "indirect" and the
English as "direct"; but when English is compared with Hebrew, it is the
English speech patterns which are "indirect" and the Hebrew "direct" (cf. for
example, Mizutani and Mizutani 1987; KatrieI1986). There is no salvation
Why are these scripts better than saying, for example, that Anglo culture
values "personal autonomy" and that in Japanese culture one has to "apolo-
gise" a lot? There are several interrelated reasons which I will adumbrate
here but not fully justify, for reasons of space. First, cultural scripts are
more easily tested and refined than are generalisations framed in technical
terms, because they are more precise and predictive, and because they are
more directly accessible to the intuitions of ordinary native speakers. Second,
unlike complex terms such as "autonomy", "apology", and "sensitivity",
which lack exact equivalents in most non-English languages, cultural scripts
composed in simple terms can be readily translated between languages,
which makes them more practical for research across linguistic barriers.
Third, because cultural scripts are framed in the same metalanguage as can
be used for lexical semantics, they provide enhanced opportunities to demon-
strate links between discourse practices and sociocultural values, as embodied
in cultural key words, value terms, proverbs, etc.
The cultural scripts approach is relatively new, and, as far as I know, it has
not been systematically applied or tested in language teaching (though it has
been used, with success, in some academic courses on intercultural commu-
nication). Nevertheless, there are reasons to think that cultural scripts could
provide significant pedagogical advantages in relation to the problem of
"bridging the culture gap" in language teaching.
In communicative and functional approaches to language teaching, and
in the emerging paradigm of Intercultural Language Teaching (Lo Bianco,
Liddicoat and Crozet, eds. 1999; Liddicoat and Crozet, eds. 2000), one of the
key tasks is to bring the learner to an understanding of those L2 concepts
and values which form the interpretive frame within which L2 individuals
negotiate meaning. In pursuit of this goal, a variety of ingenious pedagogical
techniques have been devised, which may involve analysis of texts, dia-
logues, and routines, use of role-plays, interviews, student projects, and other
activities (cf. for example, Valdes, ed. 1986, Kramsch 1993). It has often
been observed that cultural analysis in the classroom inevitably assumes a
contrastive dimension, as students become reflectively aware of how their
own concepts and values compare and contrast with those of the L2 culture.
According to a recent review of the field of interlanguage pragmatics
(Kaspar and Schmidt 1996), there has so far been little research into how
non-native speakers develop pragmatic competence, and even less into the
Though Wierzbicka (1997) was the first to approach cultural key words using
a rigorous method of semantic analysis, their existence and importance has
been recognised by numerous scholars of culture studies. Essentially, cultural
key words are conceptual "focal points" for entire cultural domains, such
that studying them leads into a dense complex of cultural values, attitudes,
and expectations. With reference to Russian examples, Wierzbicka writes as
follows:
ball of wool: by pulling it, we may be able to unravel a whole tangled "ball"
of attitudes, values, and expectations, embodied not only in words, but also
in common collocations, in set phrases, in grammatical constructions, in
proverbs, and so on ... (Wierzbicka 1997: 17)
Various grammatical facts are consistent with this explication. For example,
one can speak offreedom OF (e.g.,freedom of action, of opinion, of trade)
orfreedom TO (e.g.,freedom to emigrate), referring to something which is
desirable. One can also speak offreedom FROM (e.g.,fi-eedom from persecu-
tion, from harassment, from coercion, from external control), referring to a
and the like. These constructions clearly convey the message 'I want you to
do this', but at the same time they embed the potentially confronting message
into a question-form, with the effect of inviting the addressee to say whether
or not he or she will comply, thereby acknowledging, in form at least, the
addressee's autonomy. Part of acquiring mainstream Anglo-American
English as a child (or learning it as an adult) is internalising a personal
script like (4).
The very fact that tag questions have come to play such a major role in
English seems to reflect the same cultural attitudes which have led to the
expansion of interrogative forms elsewhere, and to the restrictions on the
use of the imperative, the same emphasis 011 possible differences of opinion.
o.lpoint olview. (Wierzbicka 1991: 40; emphasis added)
... Polish cultural tradition does not foster constant attention to other people's
'voices', other people's points of view, and tolerates forceful expression of
personal views and feelings without any consideration of other people's
views and feelings. In fact, the basic Polish tag, prawda? 'true?', presents
the speaker's point of view not as a point of view, but as an objective 'truth';
and it doesn't seek agreement but an acknowledgement of this 'truth'.
(Wierzbicka 1991: 40)
These examples (the profuse use of interrogative directive fonnulas and tag
questions in English, the preference for "hedging" in opinion-giving, and so
on) illustrate the point that cultural values can underlie even the minutiae of
everyday fonns of talk, with the implication that, as Liddicoat and Crozet
put it:
[E]ven very simple language offers genuine opportunities for cultural under-
standing in the language classroom. In fact, it is often this very simple for-
mulaic language which is most culturally loaded. (Liddicoat and Crozet
2000: lO)
Of course it would be wrong to assume that all varieties of English share the
same cultural scripts and associated repertoires of fonnulas, routines and
characteristic phrasings. For example, it can be argued that American English
Moving now to cultural scripts, I would like to begin with the claim that
Malay culture encourages concern for careful and premeditated action.
Expressions such asfikir dulu 'think first' andfikir panjang 'think long' are
heard daily. Impulsive or wilful actions (ikut suka hati 'doing what you feel
like') are viewed as dangerous and immature. As the saying goes: lkut rasa
binasa, ikut hati mati 'Follow (your) feelings suffer, follow (your) heart
die'.
Script (8) takes this a step further, spelling out the cultural priority placed on
verbal caution and premeditation, particularly in relation to hurting the feel-
ings of others: "[Malay values include] showing consideration and concern,
anticipating the other ... and, above all, being sensitive to the other person"
(Wilson 1967: 132); cf. the expressionsjaga hati orang 'watch over other
people's feelings', memilihara perasaan 'look after feelings'. The motives
here are not purely altruistic, in virtue of the strong cultural theme that people
are likely to retaliate, possibly in subtle and calculated ways (cf. the concept
of dendam 'revenge, payback'), against any offence to their maruah 'dignity,
standing' or nama 'reputation'. As the saying goes: Rosak badan kerana
mulut 'The body suffers because of the mouth'.
Needless to say, the complex of attitudes spelled out, in part, in scripts (7)
and (8) are consistent with the status of sabar 'patient, forbearing' as a cul-
tural key word.
The second key word I want to consider, hormat (verbal form: menghor-
mati), is usually glossed as 'show respect' or as 'proper politeness'. One
sociologist (Kessler 1992: 148) has described it as "deference that is owed to
a social position", stressing its role as a focus of cultural elaboration. As testi-
mony to the orthodoxy of these attitudes, another sociologist Geert Hofstede
(1997: 28) found Malaysia to be the country which ranked highest of 53
countries and regions on the so-called "power distance index", which indi-
cates the extent to which the less-powerful individuals in societal roles and
relationships expect and are willing to obey authority. (For comparison, the
USA comes in at rank 40 on this scale, and Australia at rank 36.)
Hormat involves being on one's "best behaviour", including speaking in
a nice tone of voice, not saying too much, not opposing or contradicting.
Linguistic etiquette is also very much an issue: avoiding the pronouns aku
'1', kaulawak 'you', use of various honorific words, and as far as possible
using a refined (halus) speech style. According to explication (9), the idea
behind menghormati is to show the "respected" person that you recognise
his or her higher standing and that you want to avoid his or her disapproval.
To this end, you will behave in a deliberately selective fashion - in terms of
what you do, what you say, and how you say it. 7
The phrasing of this script goes beyond vague and English-based generali-
sations about "respecting" older people, 8 while at the same time linking
closely with the explication of menghormati. Taken together, the script and
the lexical concept account for the self-evident status (from a Malay per-
spective) of the injunction to menghormati orang tua 'respect old people'.
I hope that this brief review of some cultural key words and related cultural
scripts in English and Malay has illustrated some of the points made in the
introductory sections of this paper; in particular, that the cultural scripts
approach lends itself to an integrative treatment of discourse style which
draws connections between broad cultural themes, key lexical items, proverbs
and common sayings, linguistic routines, and many other things besides
(such as child-raising practices, to mention only one matter not touched
upon in this paper). As such, the scripts approach has obvious potential for an
integrated pedagogy of pragmatics and culture. I hope also to have illustrated
the advantages of the cultural scripts approach over conventional descriptors
such as "indirectness", "politeness", "sensitivity", "respect", and so on.
Cultural scripts are clearer, more articulated, and potentially much more
accurate than vague and language-specific labels of this kind.
Because the same non-ethnocentric metalanguage can be used both for
cultural scripts and for explicating the meanings of cultural key words - the
Substantives: Time:
I, YOU, SOMEONE, SOMETHING, WHEN/TIME, NOW, BEFORE, AFTER,
PEOPLE, BODY A LONG TIME, A SHORT TIME, FOR
SOMETIME
Notes
1. The parts of this paper which are devoted specifically to Malay are based on
Goddard (2000,2001). I would like to thank Norlinda Hasan, Vicki Knox, Lee
Mee Wun, Mohamad Mokhtar Hassan, and Anna Wierzbicka for helpful com-
ments and information with these studies. For comments specifically on this
paper I would like to thank two anonymous reviewers.
2. For a number of detailed studies of the proposed metalanguage in a sample of six
genetically and geographically disparate languages see Goddard and Wierzbicka,
eds. 2002. In a previous set of smaller studies (Goddard and Wierzbicka, eds.
1994), sixteen languages were surveyed, only one of which was Indo-European.
A full description of the Malay version of the natural semantic metalanguage
can be found in Goddard (2002b).
3. It is, of course, important to recognise that cultural norms are not static or ahis-
torical: they change, interact, and vary in space and time. In order to be able to
understand this variation, however, we must first be able to describe "local"
norms with rigour and precision, without getting bogged down in complex,
language-specific, and culture-specific vocabulary. It is also important to
acknowledge that cultural scripts are not necessarily "binding" on individuals.
The claim is merely that they furnish an interpretive backdrop against which
individuals construct and negotiate meaning.
4. Unfortunately, in Berlin's discussion the English words freedom and liberty are
used interchangeably, which is confusing because the two words do not mean
the same (cf. Wierzbicka 1997: 132-134).
5. Freedom from X can also be used in situations when some condition (such as
hunger or poverty) prevents us from doing what we want to do. Expressions like
freedom from hunger and freedom from poverty have a political undertone
because they imply that hunger, poverty, and so on are conditions imposed by
other people.
6. Present-day Malaysia is a multi-ethnic country. In this paper I confine myself to
the "culture of communication" of the Malay people, who form the majority of
the population.
7. Explication (9) applies to menghormati when it is used about people. Like English
'respect' in this regard, the word menghormati can also be used in relation to non-
human entities such as one's country, rules, the environment, etc. A different
(related) explication would be required to cover such cases.
8. The relevant concept is not really relative age (being older) as such, but, roughly
speaking, who was born earlier: one should menghormati those who have 'lived
for some time before me'.
References
Ameka, Felix
1987 A comparative analysis of linguistic routines in two languages: Ewe
and English. Journal ofPragmatics 11(3): 299-326.
1994 Areal conversational routines and cross-cultural communication in a
multilingual society. In: Heiner Piirschel (ed.), Intercultural Communi-
cation. 441-469. Bern: Peter Lang.
Berlin, Isaiah
1969 Four Essays on Liberty. Oxford: Clarendon Press.
Durst, Uwe
2001 Why Germans don't feel "anger". In: Jean Harkins and Anna
Wierzbicka (eds.), Emotions in Cross-linguistic Perspective. 115-148.
Berlin: Moutin de Gruyter.
Goddard, Cliff
1992 Traditional Yankunytjatjara ways of speaking-A semantic perspective.
Australian Journal ofLinguistics 12: 93-l22.
1994 The meaning of lah: Understanding "emphasis" in Malay (Bahasa
Melayu). Oceanic Linguistics 33(1): 145-165.
1996 The "social emotions" of Malay (Bahasa Melayu). Ethos 24(3):
426-464.
1997 Cultural values and "cultural scripts" of Malay (Bahasa Me/ayu).
Journal ofPragmatics 27(2): 183-201.
1998 Semantic Analysis: A Practical Introduction. Oxford: Oxford Uni-
versity Press.
2000 Communicative style and cultural values - Cultural scripts of Mal ay
(Bahasa Melayu). Anthropological Linguistics 42 (1): 81-106.
2001 Sabar, ikhlas, setia - patient, sincere, loyal? A contrastive semantic
study of some "virtues" in Malay and English. Journal ofPragmatics
33: 653-681.
2002a Directive speech-acts in Malay: An ethnopragmatic perspective. In:
Les Cahiers de Praxematique 28 (Special Issue 'Language, discours,
culture', edited by Christine Beal), 113-143.
2002b Semantic primes and universal grammar in Malay (Bahasa Melayu).
In: Cliff Goddard andAnna Wierzbicka (eds.), Meaning and Universal
Grammar - Theory and Empirical Findings, Volume I, 87-172.
Amsterdam: John Benjamins.
Goddard, Cliff and Anna Wierzbicka
1996 Discourse and Culture. In: TeunA. van Dijk (ed.), Discourse: A multi-
disciplinary introduction. 231-257. London: Sage Publications.
Wierzbicka, Anna
1972 Semantic Primitives. Translated by Anna Wierzbicka and John
Besemeres. FrankfurtlM.: Athenaum Verlag.
1991 Cross-Cultural Pragmatics. Berlin: Mouton de Gruyter.
1992 Semantics, Culture and Cognition. Oxford: Oxford University Press.
1994a "Cultural scripts": A semantic approach to cultural analysis and cross-
cultural communication. In: L. Bouton and Yamunu Kachru (eds.),
Pragmatics and Language Learning. 1-24. Urbana-Champaign:
University of Illinois.
1994b "Cultural Scripts": A new approach to the study of cross-cultural
communication. In: M. Piitz (ed.), Language Contact and Language
Conflict. 67-87. Amsterdam: John Benjamins.
1996a Semantics, Primes and Universals. Oxford: Oxford University Press.
1996b Contrastive sociolinguistics and the theory of "cultural scripts":
Chinese vs. English. In: M. Hellinger and U. Ammon (eds.),
Contrastive Sociolinguistics. 313-344. The Hague: Mouton.
1997 Understanding Cultures Through Their Keywords. Oxford: Oxford
University Press.
1998 German "cultural scripts": Public signs as a key to social attitudes
and cultural values. Discourse & Society 9(2): 241-282.
1999 Emotions across Languages and Cultures: Diversity and Universals.
Oxford: Oxford University Press.
2002 Australian cultural scripts - 'Bloody' revisited. Journal ofPragmatics
34(9): 1167-1209.
Wilson, Peter J.
1967 A Malay Village in Malaysia. New Haven: HrafPress.
Ye,Zhengdao
2001 An inquiry into "sadness" in Chinese. In: Jean Harkins and Anna
Wierzbicka (eds.), Emotions in Cross-linguistic Perspective. 361-406.
Berlin: Mouton de Gruyter.
2002 Different modes of describing emotions in Chinese: Bodily changes,
sensations, and bodily images. Pragmatics and Cognition 10(1/2):
321-356.
Michel Achard
1. Introduction
Because its main concern is the description oflanguage in use (that is, in the
full richness of its psychological and social context) the cognitive linguistics
movement has from its beginnings been interested in the relation between
linguistic description and language learning. The very foundations of the
movement make it well suited for that task (Taylor 1993). For example, the
role of meaning in determining the form of grammatical constructions pro-
vides an intuitively appealing way of teaching those constructions. As we
shall see later in this paper, the knowledge of an expression's meaning and
communicative function greatly facilitates the design of activities to present
it in class. Also, the constructs used in the various cognitive linguistics
models (for example schemas, metaphors, idealized cognitive models,
blends) are socially and culturally rich, and thus lend themselves well to
cross-cultural comparisons that prove useful in the foreign language class-
room. However, in the absence of a fully developed cognitive linguistics L2
acquisition model, the movement's contribution to language pedagogy will
not reach its full potential if cognitive linguists do not incorporate their
descriptive insights into well-established models of acquisition and peda-
gogy.
Second language pedagogy is a truly multidisciplinary endeavor, because
any activity proposed in the classroom (a matter of pedagogy) makes an
implicit assumption about the way in which the target system is organized
(a matter of linguistics), as well as how the target construction is learned (a
matter of acquisition). However, despite the collaboration one would hope
to find given the dependence of each one of those fields on the others, just
the opposite obtains. Each discipline remains surprisingly separate from the
Il III
VanPatten posits three kinds of cognitive processes that account for the pro-
cessing of L2 structures. The first set of processes transforms the input
(what the learner hears) into intake (what she understands). The second set
of principles organizes the intake with respect to the learner's developing
system in L2 (her emergent grammar), and the third kind of processes
extract the appropriate elements of that grammar for specific conversational
purposes. We can use the model in Figure 1 to show precisely how CG can
contribute valuably to L2 acquisition and pedagogy. Let us first note that if
scope of the paper is therefore exclusively contained within the area ofped-
agogy.
Filter
Language
Acquired
Input Acquisition
Competence
Device
Figure 2. Operation of the "Affective Filter" (From Krashen and Terre1l1983: 39)
This diagram is interesting because it contains one of the very few references
to a Generative model with the mention of the Language Acquisition Device
(Chomsky 1965). The Chomskyan orientation of the NA acquisition model
ING (progressive)
PLURAL
COPULA (to be)
1
AUXILIARY (progressive)
ARTICLE (a, the)
1
IRREGULAR PAST
r
REGULAR PAST
III SINGULAR (-s)
POSSESSIVE (-s)
4. Cognitive Grammar
In this section, I briefly present the aspects of the CG framework that make it
directly compatible with the view of L2 pedagogy defended by the Natural
Approach, and thus enables it to provide a satisfactory frame for communi-
cative grammatical instruction.
4.1. Construal
In previous work (Achard 1993, 1996) I argued that the two constructions
have their own specific semantics, characterized by the specific profile they
each impose on a dynamic scene. The choice between the constructions
depends on whether the causee (John) is presented as the energy source that
initiates the process coded by the infinitive in the complement. If John is
presented as the initiator of the leaving, VOV (illustrated in 1) is chosen. If
John is not construed as the initiator of the infinitival process, VV (illus-
trated in 2) represents the favored choice.
The notion of construal is extremely important for the purposes of this
paper. In Section 5, I will consider in detail how this notion critically impacts
the teaching of grammar. For now, I will simply note that since the meaning
of a construction is a matter of conventionalized choice, its distribution in dis-
course is determined by the speaker's decision rather than by the system itself.
SCHEMA
~0
0CIJ QUAGMIRE
quagmire
§J0
~G
~
REE
I'
tree
BjL
,
-8
11
and less structured situations in the target area. I propose that grammatical
activities simply follow the activities phase within each targeted semantic
domain. This allows grammatical instruction to benefit from all the acquisi-
tion that took place in the beginning of the lesson. Since the input phase, the
students have been exposed to specific instances of the targeted grammatical
constructions, but they have not focused on their commonalties to extract
patterns of generalization. Grammatical focus is thus most appropriate when
the students are comfortable with the semantic field of the lesson and can
concentrate on extracting patterns of regularity from that field.
The transition between the activities phase and the grammar phase should
be so smooth that even though the instructor needs to work at it very hard,
the students should not even realize they are doing grammar. In fact, gram-
matical activities are communicative activities with one added particularity.
Because the meaning of grammatical constructions is more abstract than
that of lexical items, grammatical activities should be more repetitive and
focused than other activities. The learners need to be exposed to more
instances of the structure in order to learn it, and they also need more
focused activities to successfully associate the construction's form to its
meaning. This represents the most difficult aspect of grammatical instruc-
tion. The activities need to be naturally repetitive and focused. The idea is to
find situations where the repetition and focus are both natural and functional.
Let us take an example to illustrate the principles of communicative
grammatical instruction proposed in this paper. One of the first grammatical
points considered in the NA French textbook Deux Mondes (Terrell et al.
2001 Chapitre 1) concerns the generalization that some verbs, most notably
the verbs of preference such as aimer 'like', detester 'hate', or prejerer
'prefer' are directly followed by an infinitive (J'aimejouer 'I like to play').
The context for this presentation (the theme of the lesson) is Mafamille et
moi 'My family and myself', where learners are exposed to new vocabulary
and structures concerning the hobbies and pastimes families enjoy together.
Consistent with the NA practice, no grammar is presented in the course of
the lesson proper, but a special grammatical section in English (in different
colored pages) is cross-referenced with the relevant activities. This format
sets up grammatical instruction as a side activity, its remote physical loca-
tion clearly indicative of its peripheral role within the method.
By contrast, in the kind of methodology proposed in this paper, gram-
matical instruction needs to be fully integrated in the lesson. In fact, it is
important to know precisely what the students have already acquired from
other parts of the lesson to know both what the grammar presentation can
assume as background knowledge, and what it needs to concentrate on. The
activities phase, in particular, provides very good preparation for the stu-
dents' later focus on form. For example, in Chapitre 1 activity 5, (Terrell et
al. 2001: 50), one student reads a statement expressing her like or dislike for
specific activities, and another student agrees or disagrees with that state-
ment by saying oui or non. The model for the activity goes as follows.
Student 1: Pendant les vacances,j'aime voyager 'during the vacation I like
to travel'. Student 2: Moi aussi,j'aime beaucoup voyager 'I like to travel a
lot too' (Moi, non! Je n 'aime pas voyager 'I don't like to travel'). The goal
of this activity is for students to acquire the meaning of the activities rather
than the form of their expression. However, they are obviously exposed to a
lot of "grammar". The narrow range of the activity allows for repetition,
and the infinitival forms are clearly available for students to infer their gen-
eralization. The only missing element is some form of emphasis placed on
that generalization. 3 By placing the students in situations where they prac-
tice individual instances repetitively, such activities provide the necessary
background for the grammatical activities where the focus is on what those
instances have in common.
After going through a series of similar activities, and prior to the gram-
matical presentation of the V+INF rule, the students are able to understand
almost everything that relates to likes and dislikes, generate some likes and
dislikes, and recognize the forms of aimer 'like'. The goals of the grammat-
ical activity are thus to i) bring to their attention the fact that some verbs are
directly followed by an infinitive, ii) show that French infinitives are of 3
kinds (-er, -ir, -re), and iii) show that most verbs have an -er infinitive. 4
In order to achieve these goals, the grammatical activities need to be nat-
ural, focused, and repetitive enough to favor the emergence and entrenchment
of the three infinitival schemas, but nonetheless involve a valid communica-
tive task. 1 propose the following as a possible example, simply because it
has already been used with success in beginning classes. The details of the
activity do not matter much. Language instructors possess the necessary
resourcefulness and creativity to design their own activities. What is more
important is the spirit in which the activity is created, as well as the mindset to
incorporate similar ones in the flow of the NA class. The instructor "plants"
a backpack in the classroom filed with objects chosen to evoke specific
activities. For example, a book evokes reading, a tennis ball evokes playing
tennis. At the relevant point of the lesson (following the communicative
activities where the students have practiced different likes and dislikes), the
instructor notices the backpac~ and asks whom it belongs to. When nobody
claims it, she opens it to find out if it contains a clue that would identify its
owner. Obviously no such cue exists, but the objects inside indicate what
the owner does (and can thus be assumed to like). The instructor pulls out
the objects one by one, and the students indicate what the unknown person
likes to do. For example, when being shown a running shoe, they usually
volunteer elle aime courir 'she likes to run', or elle aime faire du sport 'she
likes to play sports'. Student production is then written on the board in three
columns that clearly indicate the relevant morphological groupings. The
length of the -er column shows how prevalent that pattern is in French. At
this point, additional explicit instruction exclusively depends on the instruc-
tor's taste. If deemed necessary, it is quite easy to briefly comment on the
patterns on the board in French to further focus the student' attention on the
formal aspects of their own production. This grammatical activity is obvi-
ously quite similar to the other activities proposed in class daily. However,
its repetitive nature, narrow scope (it's all in the bag!), and order of presen-
tation allow the students to focus on form, which contributes to the entrench-
ment ofthe V+INF schema and its -er, -ir, and -re instantiations.
Beyond this particular example, the practice of grammatical instruction
as it is defended in this paper involves three steps. First, the selection of a
particular point that deserved to be shown in a grammatical light. Secondly,
the design of an activity that will enhance the form of the targeted construc-
tion. Finally, the planning of the sequence of activities, so that grammatical
instruction most efficiently draws on the practice provided by both the input
and activities phase.
To briefly summarize this section, we can note that communicative
grammatical instruction impacts the NA lesson in a minimal way, because it
makes use of similar kinds of activities instructors routinely use to teach
lexical items. Grammatical activities are simply more narrowly focused and
naturally repetitive than the others, and their strategic integration in the
course of the lesson allows the students to focus on the relevant patterns
most efficiently. From this perspective, grammatical instruction does not
involve the presentation of a different domain of knowledge. The students
simply consider the same semantic domains they are exposed to in the
course of their communicative syllabus through a narrower lens. This is, of
course, consistent with the CG position that the rules of grammar cohabitate
with their specific instances in our linguistic knowledge.
which learners approach the rules of a foreign language. Recall from Section
four that linguistic production is primarily a matter of speaker choice or
construal. The centrality of construal makes strong implications for teach-
ing. It is not the system per se that is responsible for the form of specific
constructions just by being in a certain way, but the shape of novel expres-
sions is detennined greatly by speaker choice. In that sense, the study of the
mechanisms of that choice belongs to the access processes (level HI) pre-
sented in Figure 1.
The centrality of the speaker to the form of the utterance seems trivial.
However, it is important for pedagogical purposes. Focusing on the speaker
rather than on the system represents the difference between teaching set pat-
terns of lexical associations and teaching construal, that is, the convention-
alized way of matching certain expressions to certain situations, as well as
the flexibility of using the available alternatives to express specific semantic
nuances. This difference is significant for the instructor. Grammatical rules
traditionally given in a language class are considered a property of the sys-
tem, and not a result of the speaker's choice. As an example, learners of
Romance languages are taught that the subjunctive mood follows a specific
list of verbs. Beside its specific morphological shape, learning the subjunc-
tive mood therefore amounts to learning the verbs it occurs with. The fact
that a given verb takes the subjunctive is viewed as a lexical property of that
verb. Selectional restrictions are thus marked in the lexicon, and the stu-
dents learn them at the same time as the verb's meaning. Such rules are
deeply entrenched in our pedagogical history. They are also easy to state (if
not to follow) because they yield a stable system of co-occurring expres-
sions that can easily be apprehended through memory. Difficulties only
arise when learners become aware of the numerous exceptions that creep
into the supposedly regular system. Teaching construal, on the other hand,
is much harder because of its inherent flexibility. It is methodologically dif-
ficult to provide clear guidelines that teach people how to exercise choice.
In the case of a foreign language, in addition to providing the students with
the whole range of conventionalized options, instructors need to be able to
show them the precise conditions under which natives would use a particu-
lar construction.
In order to illustrate the differences between a model of lexical associa-
tion, where the constructions are considered part of the system, and the con-
strual-based approach recommended by CG, Twill now consider the case of
the distribution of definite and partitive articles in French. In addition to the
definite le, la, les 'the', and indefinite un, une, des 'a', French has a series of
partitive articles du, de la, which are often translated in English by 'some',
To choose the appropriate article, look at the kind of verb used in the sen-
tence. With verbs describing likes or dislikes, such as aimer, adorer,
detester, preferer, use the definite article because you are talking about
things in a general sense, Nathalie aime beaucoup les carottes et les petits
pois, mais elle deteste les epinards 'Nathalie likes carrots and peas, but she
detests spinach', Je n'aime pas le cafe fort 'I don't like strong coffee'.
On the other hand, if the verb deals with having, obtaining, or consuming,
use du, de la, de l', or des, because you are talking about some amount of a
thing. Such verbs include avoir, acheter, manger, boire, prendre, and
many others. Les Fran~ais boivent du cafe apres le diner 'the French drink
coffee after dinner'. Nous mangeons de la pizza tous les vendredi soir 'We
eat pizza every Friday night'.
tous les matins with a slight difference in meaning to which we will come
back later in this section. This is particularly important in a method such as
the NA that places such strong emphasis on input. The instructor is expected
to use the widest possible range of comprehensible expressions in her
speech in order to enrich the students' understanding of a given domain.
When talking about drinks and socializing, the students are thus likely to
hear bothj'aime boire du cafe andj'aime boire le caFt Admittedly, they
will not experience a total breakdown in comprehension, but they may be
confused, or simply miss the cultural dimension of the meaning distinction
between the two forms.
The reality of usage is such that in numerous cases, the expected co-
occurrence of the consuming verbs with the partitive on the one hand, and
the liking verbs with the definite article on the other hand, is violated. The
following are a few random examples:
(5) A la cantine, hier, pour la premiere fois j'ai pris le poulet, c'etait pas
mauvais
'At the cafeteria yesterday for the first time I took the chicken, it was
not bad'
Note that in (3) and (6), the consumption verb boire is unexpectedly fol-
lowed by the definite article. In the same way, prendre is also followed by
the definite article in (5). Furthermore, some verbs such as acheter in (3)
and (4), and prendre in (7) can felicitously be followed by different articles.
Finally, partitive articles can also follow the liking verbs, as shown in (8):
This small sample of examples should suffice to show that pure lexical
association is incapable of rendering the complexity of article distribution in
French, even from the beginning. To explain the exceptions, teachers have
to backpedal from the convenient generalization expressed earlier, and try
to explain each case on an individual basis. This wastes useful class time,
often involves explanations that need to resort to the source language, and
eventually does little to alleviate confusion.
Rather than imputing the distribution of mticles to the main verb, it seems
more appropriate to explain it in terms of the way in which the entity evoked
by the verbal complement is construed. The CG account of article distribution
involves the recognition of the meaning of both finite and partitive articles,
characterized by the way in which they relate the object nominal to the
speech situation. It is far beyond the scope of this paper to propose an exact
characterization of either of the articles, but some broad generalizations will
suffice to give the flavor of the CG analysis and how its adoption would
alter current teaching practices in the Natural Approach.
The French definite article marks generic nouns (as in J'aime le cafe), as
well as nouns which are judged identifiable by the hearer. With the partitive,
the identifiability of the nominal is not at issue; the focus is on the fact that
it represents a certain quantity of the substance referred to by the nominal
(Je bois du cafe). Competition between the two articles arises when a given
situation can be construed in two alternative ways, i.e., when a nominal can
be considered with respect to its identifiability, or with respect to its mere
quantity. In certain contexts, even with the verbs that usually favor the con-
strual of the nominal as a quantity, some discourse conditions render an
alternative construal possible, where the identifiability of the nominal is
emphasized. Generally, the referent of the entity has already been men-
tioned, or it is obviously available to the hearer as part of her background
knowledge. For example in (3), because beer and wine are mentioned first,
they can then be treated later as identifiable entities, i.e., the ones that were
just recently mentioned. Similarly in (6), the contents of the refrigerator is
part of the roommates' background knowledge. They can both easily identify
each item in it. Although the person who drank the beer and the wine neces-
sarily drank a specific amount of both, the emphasis in the utterance can
therefore be placed on the items in the refrigerator as identifiable entities.
The quantity consumed is irrelevant; what matters is that both entities have
now disappeared. In a different discourse context, where the focus is on the
quantity of liquid ingested, the partitive would constitute a more appropriate
choice. The case is similar in (5). Although customers in restaurants are
served a certain quantity of chicken, the chicken is not construed as a quantity
of a specific animal but as an item on a restaurant menu, and thus identifiable
within that schema.
It is therefore clear that the data in (3)-(8), as well as other numerous
examples not mentioned here, do not constitute exceptions to a rule of co-
occurrence. Rather, they illustrate cases where despite the lexical semantics
of the main verb, some aspect of the situation favors the construal of the
nominal as an identifiable entity. The pedagogical challenge consists in
teaching those situations, so that, when students are placed in the appropriate
situation, they can make the same choices the natives make, and enjoy the
same flexibility of expression. From a pedagogical standpoint, there are two
aspects of construal that need to be treated separately, especially at the begin-
ning level of instruction. The first one concerns the linguistic choices that
are socially conventionalized and therefore stable. The second one concerns
the more flexible, on-line choices made by specific speakers as they react to
the particularities of a given situation. Since my focus here is on the first
year of instruction, I will mostly be concerned with the first aspect.
The basic idea behind the teaching of construal is that students should be
placed in situations where native speakers are the most likely to exercise a
specific choice. The best place to start therefore is where that choice is
detennined by the presence of a visible cultural schema. Cultural schemas
constitute strong pedagogical assets because they are both meaningful
(therefore culturally rich) and easily teachable (see in particular Goddard
this volume). As an example, let us use the utterance in (7a). Prendre le cafe
is an unexpected collocation because the consumption verb prendre is
expected to be followed by a partitive article. However, the expression
refers to much more than just drinking coffee. It evokes a specific part of a
large food and drink schema in which people drink certain beverages at dif-
ferent times of the day. This ritual does not only involve coffee. The other
beverages that also mark the social rhythm of the French day also appear in
the same Def. Art -I- Noun construction. Combinations such as prendre
l' aperitif' take a drink before lunch!dinner' , prendre le digestif 'take an
after dinner liquor', prendre le the 'take a 5 o'clock snack' almost have the
status of fixed expressions. They are limited in number, and not particularly
productive. For example, *prendre le jus d'orange 'take the orange juice',
or *prendre le vin 'take the wine' are infelicitous in that sense, although
these forms would indeed be acceptable if the speaker were reading off a
grocery list.
Communicative activities can be created to link these apparently excep-
tional uses of the definite article to the culturally marked episodes in which
they appear. Consequently students will associate the whole expression
prendre fe cc~fe to a specific social act, and not analyze the verb and article
combination as a confusing sequence. Successful activities will focus on the
context of the drinking episode, namely, the rhythm of a person's day and
the social circumstances of the drink, and not on what is actually consumed.
F or example, a guessing game can be designed with two students working
in pairs. One of them reads off a specific set of circumstances that surround
the social activity someone is involved in, without mentioning any bever-
age. The other student guesses what that activity is. A possible sequence
could be as follows. Student 1: Mt: Dumas est au bar. 11 parle avec ses amis.
Il est 11 heures et demie du matin 'M Dumas is in a bar, he is speaking with
his friends, it is 11: 30 a.m.' Student 2: If prend ['aperitif 'He is taking a
drink before lunch'. 5
The same kind of analysis can be given for situations where the unex-
pected use of the definite article follows a precise schema, such as a restau-
rant menu or a shopping list, where prendre le + noun is also found, but
marks a different situation. Activities can be designed where menus are
passed to students who need to order from a set of dishes, hence justifying
the use of the definite article. Note that in such tight context (and yet mean-
ingful if the activity is done properly), an answer such as C'est sur la liste
'It's on the list' to a student's query as to why la and not de la is used
constitutes a legitimate grammatical explanation. I will not present a whole
lesson here, but the idea is to associate a particular linguistic usage to its
communicative function to show that conveying that function represents the
meaning of the construction. Attention to meaning necessarily entails atten-
tion to form, because students need to have access to the proper form in
order to convey the intended meaning. Teaching the whole range of a con-
struction is accomplished by targeting each meaning separately, and inte-
grating the activities designed to enhance its awareness to the general flow
of the NA lesson.
As indicated earlier, this kind of teaching does not address the other
aspect of construal, namely, the individual flexibility speakers enjoy in their
6. Conclusion
Notes
1. In the CG view of acquisition, the different processes are not as clearly delin-
eated as they are in Figure 1. For example, we will see that construal is at the
same time crucial for the way in which a system is organized (level IT), as well
as for the selection of a given structure in a given situation (level TTT).
2. Note importantly that the adoption of the CG model can easily account for the
facts that the NA hypotheses are attempting to capture. For example, even
though the distinction between acquisition and learning is not compatible with a
CG view of language, it is quite possible that different levels of entrenchment
and abstraction of their meaning allow different constructions to be learned
differently. The precise way in which each construction is learned must be
investigated in a rigorous acquisition research program. Similarly, CG would
not posit an existence of a Monitor, but different constructions may well benefit
from direct explicit instruction at different levels, depending among other things
on their own entrenchment, generality, and clarity of meaning. Finally, the order
of acquisition of specific morphemes constitutes an empirical issue that needs to
be confirmed by multiple studies in different languages.
3. Some NA practitioners would argue that this exposure is sufficient, that no further
exposure on form is desirable, because the students will extract the necessary
generalizations in due time. This paper contends, however, that specific gram-
matical activities can help the students by bringing into focus the generalized
patterns that constitute grammatical rules, while preserving the spirit of the NA
approach.
4. The systematic presentation of the -er verb paradigm should be done in a separate
activity.
5. The success of this activity depends on its integration in the general course of the
lesson. The considerations presented in Section 5.1 obviously also apply here.
References
Achard, Michel
1993. "French causative constructions, word order followingfaire,laisser,
andforcer." In Proceedings of the 19th Berkeley Linguistics Society.
Berkeley, University of California: 1-12.
1996. "Two causation/perception constructions in French". Cognitive
Linguistics 7: 315-357.
1997. "A Cognitive Grammar view of second language acquisition." Joumal
ofIntensive English Studies 11: 157-177.
Bates, Elizabeth, and B. MacWhinney
(1981). "Second-language acquisition from a Functionalist perspective:
Pragmatic, semantic, and perceptual Strategies". In H. Wnitz (ed.),
Native Language and Foreign Language Acquisition. Annals of the
New York Academy of Sciences, vo!. 379. New York, New York
Academy of Sciences: 190-214.
Cadierno, Teresa
This volume. "Expressing motion events in a second language: A cognitive
typological approach".
Chomsky, Noam
1965. Aspects ofthe Theory of Syntax. Cambridge, Mass.: MIT Press.
Dulay H, and M. Burt
1977. "Remarks on creativity in language acquisition". In M. Burt, H. Dulay
and M. Finocchiaro (eds.), Viewpoints on English as a Second
Language. New York: Regents: 95-126.
Elman, Jeffrey
1990. "Finding structure in time". Cognitive Science 14: 179-211.
Gasser, Michael
1990. "Connectionism and universals of second language acquisition".
Studies in Second Language Acquisition 12: 179-199.
Goddard, Cliff
This volume. "Cultural scripts: A new medium for metapragmatic instruction?"
Herron, Carol and Michael Tomasello
1992. "Acquiring grammatical structures by guided induction." The French
Review 65: 708-718.
Krashen, Steven
1981. Second Language Acquisition and Second Language Leaming.
Oxford: Pergamon Press.
Krashen, Steven, and D. Terrell Tracy
1983. The Natural Approach: Language Acquisition in the Classroom. San
Francisco: Alemany Press.
Langacker, Ronald W.
1987. Foundations of Cognitive Grammar. Vo!. 1: Theoretical Prerequisites.
Stanford: Stanford University Press.
Angeliki Athanasiadou
1. Introduction
3. When
(i) previous to given time reference (first I get back home and then I tell all
the news):
(1) When I get back home I'll tell you all the news. (Quirk and Greenbaum
1973:231)
(ii) simultaneous with given time reference (when you come next week we
shall be talking about solar energy):
(2) Join us next week when we shall be talking about solar energy.
(CCELD)
(3) He was already born when the war broke out. (Radden and Dirven, to
appear: 84)
(4) He went away when 1 visited her. (Quirk and Greenbaum 1973: 339)
(5) When he returnedfram work, his wife was in the kitchen. (Quirk and
Greenbaum 1973: 340)
The two states of affairs may exhibit every possible combination of time
point or time duration.
The wide occurrence of when in all time settings, in all possibilities of
representation of events on the time axis, i.e., point or duration, and in all
combinations of figure-ground reversal (the when-clause may be either pre-
posed or postposed to the main clause) make it a very useful but also a very
vague connective. Due to this wide variety of occurrence, apart from pure
time, it can extend to a wide variety of abstract domains and express contrast,
condition, reason, adversativeness.
3.1. Contrast
(6) 1 was going out when there was a knock at the do07" (CCELD)
Although the time concept is dominant, the notion of contrast between the
plan of going out, which is a durational event, and its being stopped by the
knock on the door, a time point event, is moreover evoked by the figure-
ground reversal. When introduces a subclause that contains the important
new event which happens suddenly and unexpectedly; and this is the figure,
while the main clause constitutes the ground. There are two reasons why in
(6) the contrast is highly dramatic and superimposed on the time concept.
First of all there is a clash between what is normally the ground (the when-
clause) and what happens here in the when-clause. The current analysis in a
cognitive-linguistic perspective is that complex sentences display a figure-
ground configuration, whereby the main clause is the figure and the subclause
is the ground. In purely temporal uses of when-clauses both the ground and
the figure can be preposed. But in unmarked, in terms of time, cases the
ground is given first so that the figure can be projected against it as in (6).
This explains why in cases like these the when-clause cannot be preposed
(* When there was a knock at the door I was going out). The second reason
of contrast is that a stretch of time, namely the activity of going out, is inter-
rupted by an abrupt time point event. When is then equivalent to 'and sud-
denly' and it can be paraphrased as 'I was going out, and suddenly there
was a knock at the door'. So (6) mainly has a contrastive meaning which is
due to the temporal combination of the two events, a time duration+a time
point.
The contrast is not so dramatic and it is not superimposed on the time
concept in a scene where the two events can be taken as simultaneous
points in time:
Here the temporal and the contrastive meanings go together. (7) has two
readings: a purely temporal one which can be seen if we apply the test of
the reversal When she entered we were having a chat and it can be para-
phrased as 'At the moment she entered we were having a chat' and a con-
trastive one as in (6) 'We were having a chat and suddenly she entered'.
One might, then, wonder whether we have a situation similar to (7) if we
change (6) to I was going out when she knocked at the door and, moreover,
if we can have the reversal of the events When she knocked at the door, I
was going out. Even if we change the type of the verb (from there was a
knock at the door to she knocked at the door) the meaning is still the same
because we have the succession of the two events: the event of going out is
prior to the event of knocking. However, in (7) the two events occur simul-
taneously - which explains the possibility of their reversal. Furthermore,
the difference between (7) and (6) is that a knock at the door is more
momentary than entering a house. That is why the contrast is so dramatic in
(6) - another explanation of the fact that the contrastive meaning of (7) is
milder than in (6).
The notion of contrast gets the upperhand, the time factor still being present,
but now it is further shifted into condition or reason-giving.
(8) How can I ever get a job in America when I can't remember any
English? (Radden and Dirven, to appear: 85)
(8a) How can I ever get ajob in America if I can i remember any English?
(8b) How can I ever get a job in America since I can i remember any
English?
of the two events allows a great many inferences when we interpret exam-
ples like these, which is evidence for the fact that a sentence is more than
the sum ofthe interpretations of the individual elements.
3.3. Adversativeness
(9) You describe this policy as rigid and unflexible when in fact it has
been extremely flexible. (Radden and Dirven, to appear: 85)
Language use has shown that relations between events are not spelled out
by explicitly using the connectives that are considered to represent them.
The explicit use of the temporal connective when is considered as a break
on the contrast continuum which varies from mild to adversative. So if two
events are related merely temporally, this frustrates the processing of the
information, it reduces the enrichment of the interpretation, and, moreover, it
deprives the speakers or the hearers to infer in a variety of ways and to decide
whether they prefer to leave the relation between events underspecified or
not.
So learners of English face a double difficulty: on the one hand they have
to master the concepts and the conceptual relations and on the other they
have to master the linguistic forms that can express this conceptual informa-
tion. For this reason they must be taught to infer the relationship intended as
well as to leave relations between events underspecified. This means that
they must be exposed to the wide variety of the uses of when which will
give them an indication of its vague character and of the non-iconic, though
prototypical, use of the continuum of contrast which seems to be superim-
posed on its temporal characteristics.
In sum, two simultaneous time duration events linked with when allow
many inferences and are intended as such by speakers. It is also the case
that prototypically non-sequential or both sequential and non-sequential
when can be enriched by a causal relation:
(10) I couldn't work when the television was on. (Traugott and Konig
1991: 197)
An additional reason for causality is the fact that we have a sequence of two
states. According to Traugott & Konig (1991: 197) states are more likely to
give rise to causal interpretations than sequences of events. The temporal
succession of two events, namely a time point + a time duration event leads
to an adversative interpretation; adversative settings are postposed since
they provide justifications or explanations of the events described in the
main clause.
The two conjunctions as long as and as soon as have extended their temporal
meanings into a conditional one. In its temporal meaning as soon as presup-
poses a time point:
(11) As soon as she got out ofbed the telephone stopped ringing. (CCELD)
When a second event happens as soon as a first event happens, the second
event happens immediately after the first event. In its temporal meaning as
long as presupposes time duration:
(12) You couldn't turn the heat off as long as the system was operating.
(CCELD)
(13) As soon as we get the tickets we'll send them to you. (CCELD)
(14) We were all right as long as we kept our heads down. (CCELD)
In (13) as soon as can be paraphrased by 'the time when' and 'if' and in (14)
as long as can be paraphrased by 'during the time that' and 'on the condition
that'. The time point setting and the 'immediately after' interpretation of as
soon as are more obvious when the as soon as-clause is preposed. In fact it
could be replaced by when. Both conjunctions, however, express a prospec-
tive temporality which naturally leads to the abstract use of the domain of
condition.
So the choice of the adequate linguistic form to express the relation of
condition can be seen as a skill learners have to master. Of course, the
capacity to choose a specific connective develops over the years and the
above specific connectives may be acquired at a later stage than more general
and vague ones like when. Research on first language acquisition, reported
by Bloom (1991), Bloom and Capatides (1987), among others, indicates
that specific connectives such as although are acquired at a later stage than
general ones like but or and. So older children are expected to use specific
forms, whereas younger children are expected to use general forms. The lat-
ter group uses underspecified relations more often than older children. One
might then expect second language learners, due to insufficient language
proficiency, to be explicit, i.e., not to leave relations implicit. But this is only
speculative and must be tested on the basis of a wide variety of parameters
such as age, situation of use, background knowledge oflearners. Tt is true
that the specific use of the connectives as soon as and as long as occurs
only when learners master the relations between events to such a degree that
they are able to refine their language according to their needs and the needs
of the hearers. They should, however, become suspicious of why both con-
nectives express prospective temporality which leads to conditionality. As
long as basically links two simultaneously durational events while as soon
as links two consecutive ones.
The examples met in Collins Cobuild and in Longman dictionaries have
shown that the relationship of the events described by both connectives is
future oriented. The prospective character of as soon as as well as the imme-
diate consequence or succession of another event can be paraphrased as
'from the moment when something happens something else will happen
immediately'. This naturally leads to the future hypothetical interpretation
of ~f As long as is also characterised by prospective temporality but the
durational nature ofthe events described leads to more general hypothetical,
namely, to the conditional interpretation: 'one thing can happen only on the
condition that another thing happens'.
Things are different with while, as, and since: they either express purely
temporal settings or they are used in a non-temporal, more abstract way. But
the two meanings seldom overlap. Purely temporally, while denotes two
simultaneous duration events:
(15) He stayed with Mom and me while Dad sat with Dr. Leon in the living-
room. (CCELD)
(16) While he was turning the key in the lock, someone opened a door on
the other side of the corridor. (CCELD)
Since only denotes posteriority. It means 'from the time that' and marks the
lower temporal boundary of the event in the main clause. It may also signal
an overlap with some point in an earlier event.
(20) They haven ~ met since the wedding last year. (Longman)
Both while and as denote two duration events that occur simultaneously.
However, one cannot replace the other: (15) *He stayed with Mom and me
as Dad sat with Dr. Leon in the living-room, or (17) *While time passed,
things seemed to get worse. This is due to the fact that as is conceptually
more delimited or bounded (e.g., (19» than while. Moreover as can also
indicate two consecutive time point events (19), whereas while always pre-
supposes at least one duration event which, as a result, makes it less
bounded than as. Since is far more specific than both as and while because
it implies a duration event after a specifically given time point.
(21) While the application may not be directly in your area of expertise,
your assessment would be appreciated.
(22) So while I have sympathy for these fellows who reacted against the
formality of their predecessors, I think they went too far. (CCELD)
So with while when two stretches of events occur simultaneously they are
contrasted. The same happens with as where the two simultaneously dura-
tional events extend to adversativeness:
(23) Flattered as I was by his attention, I somehow knew that he wasn ~ the
man for me. (CCELD)
So mere simultaneity of events is hardly relevant unless the two events are
counter to expectation and extend to concession or adversativeness. We can
see this with other connectives as well, e.g., and, an explicit additive con-
nective which has a concessive reading:
(24) He can play the Beethoven sonatas and he s only seven years old.
(Traugott & Konig 1991: 200)
(25) A cup of tea? 1 hardly think so as I'm going out in about two minutes.
(Longman)
(26) Since it was Saturday, he stayed in bed an extra hour. (CCELD)
The main clause indicates the consequence of the as- and the since-clause.
Stating reasons with as and since can occur before or after the main clause
either because speakers may explain why something is the case (27) or they
may introduce a reason for something especially when this reason involves
something that is already known to the person one is talking to (28).
(28) a. As you are here, why don't we discuss our plans? (Radden and
Dirven, to appear: 88)
b. The days were short since it was now December. (Radden and
Dirven, to appear: 89)
In (27) and (28) reasons seem to be pragmatic operations that give explana-
tion after the information of the main clause. They contain invitations for
action (27a, b, 28a) by addressing the hearer directly and by appealing to
old information already known to both participants of the situation (27a, b,
28a, b).
It was stated before that these abstract extensions have no temporal ele-
ments left. However, the conjunction as does not just express time (17, 18, 19)
or just reason (25, 27a, 28a) or just adversativeness (23), but a combination
of the above domains:
The two domains that overlap are time and reasoning: the simultaneous
duration of the two events, i.e., weeping and telling her story, leads to the
meaning extension of time into reason. So apart from the component of time
duration of as, the domain of reason is also present since the explanation of
why she was weeping bitterly is provided. The fact that in (29) both domains
of time and reason overlap, but especially the presence of the domain of
time, makes the replacement of as by since impossible. As was said above,
since is temporally unspecified.
Whereas simultaneous events are interpreted as having a link of reason
(29), consecutive events are interpreted as having a causal link, apart from
the temporal one:
Causality is not only due to the fact that the as-clause is preposed, which
iconically mirrors the cause-effect relationship, but also because of the
physical circumstances that surround an event: the fact that he offered to
help her, caused her crying. Thus, the general effect of consecutiveness is
causal interpretation. Events, then, that follow one another consecutively
can be interpreted as having a causal link. Here consecutiveness is addition-
ally stressed by the verb begin which explicitly marks a new event. Since
could not replace as here either, as we would get a totally different interpre-
tation.
6. Concluding Remarks
Utterances like the ones treated in this paper are rather frequent in natural
discourse, and despite the fact that they constitute instances ofunderspecifi-
cation, they do not seem to create interpretive problems. When native
speakers were asked, they were perfectly able to identify the intended hidden
meaning or the meaning that was superimposed on the temporal one. The
question, of course, remains as to how they can be taught. On the basis of
the process of metaphor, involving shifts from the external temporal to the
internal abstract domains, learners can be presented with the following sum-
marising table of the temporal connectives:
On the basis of the above table, it can be seen that two events that temporally
follow one another or occur simultaneously and are linked by the above
connectives tend to express degrees of contrast as well as degrees of causality.
This is natural since mere succession or co-occurrence of two events is rarely
highly relevant information unless the succession or the co-occurrence is
either contrary to expectation or a relation of cause and effect is implied.
These specifications are gradually incorporated into the lexicon. The situation
must be as follows: in the beginning, first or second language learners are
willing to produce and link complex events but they master few and general
temporal connectives, hence the large number of underspecified relations.
Later, when they master language sufficiently, they make a conscious
choice between explicit and implicit use of connectives. Being aware of the
needs of their addressees, learners may be advised to be explicit or implicit.
This shows that learners must become aware of or must look for contrastive
and causal connections because these two types of connections play a central
role in our mental knowledge. They must be taught to rely heavily on as-
sumptions concerning contrast and causality when processing information,
and, moreover, to be able to distinguish degrees of contrast and causality. In
particular, the contrast between two events may be mild or dramatic or may
extend to concession or adversativeness whereas causality may be objec-
tively construed and thus reflect the cause and effect link. It may also be
subjective and then reflect the reason and consequence setting. Condition-
ality is linked to causality and concession: a conditional relation can be seen
as a hypothetical variant of a causal relation, a concessive relation as an
inoperant cause. The affinity between degrees of causality and degrees of
contrast reflects their deep-lying relationship which is apparent in cognition
and discourse (Couper-Kuhlen & Kortmann eds. 2000: 2).
Learners, then, must be taught to rely on the different, more specific,
non-temporal, meanings of the above connectives which basically involve
assumptions of causality and contrast. This is facilitated when there exists a
continuum situation, i.e., when there are no breaks but the succession or
simultaneity of the described events on the temporal axis. This indicates that
the explicit use of the temporal connectives hinders or even stops the further
processing of information. So the implicit contrastive or causal readings of
the temporal connectives must be favoured and must be emphasised when
taught, because many possibilities are open for specific inferences. Psycho-
linguistic research on causality and contrast might prove useful in the way
events linked by temporal connectives are processed, in the way they are
retained and can be recalled best. But these are aims beyond the scope of the
present paper, which only points out the potential difficulties of connectors
for learners and provides some speculative suggestions for their teaching.
Practising teachers must anticipate the difficulties learners will experience
and try to find principles which will predict learning problems (Taylor 1993:
213). One such principle might be the fact that the conceptualisation and the
construal of the non-temporality of the temporal connectives is central and
prototypically lies in the learners' consciousness.
References
Bloom, L.
1991 Language development from two to three. Cambridge: Cambridge
University Press.
Bloom, L. and Capatides, J.
1987 Sources of meaning in the acquisition of complex syntax: The sample
of causality. In: Journal of Experimental Child Psychology, 43,
112-128.
Claudi, Ulrike and Bemd Heine
1986 On the metaphorical base of grammar,.In: Studies in Language lO:
297-335.
Collins Cobuild English Language Dictionary (CCELD).
1988 London and Glasgow: Collins.
Couper-Kuhlen, Elizabeth, Bemd Kortmann (eds.)
2000 Introduction. In: Elizabeth Couper-Kuhlen and Bemd Kortmann
(eds.) Cause-Condition-Concession-Contrast, 1-8. Berlin, New York:
Mouton de Gruyter.
Grice, Paul
1975 Logic and conversation. In: Syntax and Semantics Ill: Speech Acts,
Peter Cole and Jerry Morgan (eds.), 41-58. New York: Academic Press.
Lakoff, George and Mark Johnson
1980 Metaphors we Live by. Chicago: Chicago University Press.
Longman Dictionary of Contemporary English.
1995 Third Edition.
Quirk, Randolph and Sidney Greenbaum
1973 A University Grammar ofEnglish. Longman.
Radden, Giinter and Rene Dirven
to appear Cognitive English Grammar. John Benjamins Publishing Company.
Spooren, Wilbert
1997 The Processing of Underspecified Coherence Relations. In: Discourse
Processes, 24,149-168.
Sweets er, Eve
1990 From Etymology to Pragmatics. Cambridge: Cambridge University
Press.
Taylor, John
1993 Some pedagogical implications of cognitive linguistics. In: Richard
Geiger and Brygida Rudzka-Ostyn (eds.) Conceptuaiizations and
Mental Processing in Language. CLR 3. Berlin, New York: Mouton
de Gruyter.
Traugott, Elizabeth Closs and Ekkehard K6nig
1991 Semantics-Pragmatics of Grammaticalization Revisited. In: Elizabeth
Closs Traugott and Bemd Heine (eds.) Approaches to Grammati-
calization, Volume 1, 189-218. John Benjamins Publishing Company.
Frank Boers
1. Introduction
One of the central tenets ofCogmtive Semantics (e.g., Gibbs 1994, lohnson
1987, Lakoff 1987) is that the meaning of many conventional figurative
expressions, including figurative idioms, is motivated rather than arbitrary.
This runs counter to the "traditional" conception of figurative idioms as
"dead" metaphors whose meaning has become completely arbitrary. While
the "traditional" view maintains that idiomatic language is largely unsys-
tematic, Cognitive Semantics endeavours to look for systematicity in figu-
rative language. When applied to FLA (foreign language acquisition), the
fonner view implies that most idioms can only be learned through random,
blind memorisation. In comparison, the Cognitive Semantic proposal looks
more appealing, as it carries the potential of alternative learning strategies
(e.g., Boers and Demecheleer 1998; Lazar 1996; Low 1988).
These alternative strategies rest on an enhanced metaphor awareness on
the part of the language learner. Such an enhanced metaphor awareness
involves (i) recognition of metaphor as a common ingredient of everyday
language; (ii) recognition of the metaphoric themes (conceptual metaphors
or source domains) behind many figurative expressions; (iii) recognition of
the non-arbitrary nature of many figurative expressions; (iv) recognition of
possible cross-cultural differences in metaphoric themes; and (v) recognition
of cross-linguistic variety in the linguistic instantiations of those metaphoric
themes.
In an FLA context, an enhanced metaphor awareness can be applied to
the acquisition of conventional figurative expressions in roughly three
ways: the imagery behind idioms can be made explicit by refelTing to the
literal (or original) meaning; learners can be encouraged to invest cognitive
effort in trying to figure out the meaning of idioms independently; and
idioms can be grouped under common metaphoric themes. The results of
several controlled experiments have shown that these strategies can indeed
be very beneficial to learners' retention of unfamiliar figurative expressions.
The three proposed learning strategies and the related experimental findings
are outlined in the following section.
group were given the task to hypothesise about the etymological origins of
each of the idioms. In subsequent tests measuring participants' retention of
form as well as meaning, the experimental group scored consistently and
significantly better than the control group (p.<OOl).
A second way in which an enhanced metaphor awareness can be benefi-
cial to vocabulary retention exploits learners' problem-solving skills. If it is
true that the meaning of many idioms is motivated, then learners may be
encouraged to figure them out independently before turning to the teacher
or to a dictionary for help (e.g., Lennon 1998). Inferring the meaning of
unfamiliar figurative expressions requires cognitive effort and involves
deep cognitive processing, which raises the probability of memory storage
(e.g., Ellis 1994). Guessing the meaning of "imageable" idioms such as
keeping something under one s hat and crying over spilt milk appears not to
be beyond the capacities of many language learners, even at lower-interme-
diate levels of proficiency. In an experiment (reported in detail in Boers and
Demeche1eer 2001) 78 French-speaking university students were presented
with a range of unfamiliar English imageable idioms and asked to "guess"
their meaning. Despite the absence of any contextual cues, about 35 % of the
participants' responses overall were correct. Contextual cues (which are
usually available in "normal" learning conditions) obviously add to the
probability of successful inferences (e.g., Cooper 1999).
A third major application of metaphor awareness to vocabulary expansion
rests on the existence of general metaphoric themes. What I call metaphoric
themes largely corresponds to the notion of conceptual metaphors in Cogni-
tive Semantics (e.g .. Lakoff 1987). They are conventionally presented in
capital letters, while their instantiations (i.e., the actual linguistic expres-
sions) are listed in italics. For the present purposes I prefer using the tenn
metaphoric theme, because the use of the term conceptual metaphor might
carry the risk of luring us into the ongoing theoretical debate around the
exact conceptual status and cognitive functioning of metaphor. That debate is
probably beyond the scope of a chapter which desperately tries to maintain
an applied linguistics focus. For detailed information about the different
camps in the theoretical debate I refer the reader to Engstrom (1999),
Fauconnier and Turner (1998), Glucksberg, Manfredi and McGlone (1997),
Katz (1998), Murphy (1996), Vervaeke and Green (1997), and Vervaeke and
Kennedy (1996). In an SLA context, metaphoric themes may offer a frame-
work for the integration of knowledge. Since numerous expressions can
often be traced back to a common source domain, they can be presented to
learners as belonging to the same category. Grouping idioms under a shared
metaphoric theme is a way of lending structure and organisation to a world
The aim of the present article is to frame the encouraging results of those
learning experiments by (i) establishing the conditions under which raising
learners' metaphor awareness is likely to be most fruitful, (ii) defining the
role of the teacher in this process, and (iii) identifying areas that invite fur-
ther exploration. While strongly advocating the use of metaphor awareness
to expand learners' vocabulary, this article will nevertheless try to fine-tune
expectations along three dimensions: (i) the kind of expansion that we may
realistically hope for, (ii) the kind oflearners who are most likely to benefit,
and (iii) the kind of vocabulary that may most suitably be seasoned with a
touch of metaphor awareness.
A first practical concern for language teachers may be how much time and
effort should be invested in raising students' metaphor awareness in order to
get long-term returns. Could a single eye-opener suffice, or do learners con-
stantly need to be reminded of the metaphoric nature of the expressions they
encounter? For some linguists it has been sufficient to read the first couple
of pages of Metaphors we live by (Lakoff and 10hnson 1980) to perceive
metaphors everywhere for the rest of their lives. However, we can hardly
take this prolonged effect of a one-off eye-opener in all language learners
for granted.
Firstly, to my knowledge there is as yet no hard empirical evidence of
beneficial effects on vocabulary retention in the long term. In all the afore-
mentioned learning experiments, the participants were tested rather ShOlily
(within two weeks) after the novellexis had been presented to them. Secondly,
there is no guarantee that a learner's enhanced metaphor awareness when
processing one set of figurative expressions will lead that learner to transfer
this insight to her processing of other figurative expressions in the future.
The question of a potential long-term effect on vocabulary retention as well as
the question of potential transfer to new figurative expressions invite further
research.
In a very modest attempt at finding preliminary answers to these research
questions, a follow-up test was administered to the university students who
had participated in the aforementioned experiment on up-down expressions
(soar; plunge; slide; etc.), after a year had passed.
In the period between the first experiment and the follow-up, these stu-
dents of business and economics had continued their common English
courses. The programme was divided over two tenns, intemlpted by a two-
month summer break. Each term contained about 30 hours of English
tuition, given by different teachers to small groups of students. We managed
to track down 30 students who had received the experimental treatment
(i.e., a one-off eye-opener concerning the metaphoric nature of certain up-
down expressions) and 19 students who had received the more traditional
treatment. Another 9 students who had been in that control group now had
to be considered as a separate, third category, for the following reason.
During the term immediately following the first experiment, these individuals
had received tuition from a teacher (the author) with an almost unhealthy
interest in metaphor. This teacher had not been able to resist the temptation of
regularly drawing his students' attention to the metaphoric nature of various
expressions. Whenever figurative expressions in the course materials (such
as a heavy debt burden; getting back onto the right track; pruning expenses;
beingfed up; etc.) could be motivated, their meaning was clarified by refer-
ring to a concrete scene or source domain.
The follow-up task was similar to the task in the first experiment: the
participants were asked to write a short essay describing graphs. The essays
were then collected and screened for the occurrence of the up-down expres-
sions that had been studied the previous year (skyrocket; dive; climb; etc.).
The results did not show a significant difference between the range of
up-down expressions produced by the experimental group and the control
group anymore. The average numbers of different targeted expressions were
3.67 and 3.16, respectively. This means that either the experimental group
had regressed or the control group had caught up in the one-year period.
Interestingly, however, the average number of different targeted expressions
(soar; slide; etc.) used by the third group (the students who had received 30
hours of tuition by a metaphor addict) was 4.89, which did turn out to be
significantly higher (p .007). In other words, the lexical resources tapped for
this patiicular task appeared larger in students whose metaphor awareness
had been raised on several occasions.
Considering the small scale of the experiment and the many variables
involved, we can draw only very tentative conclusions from these findings,
of course. It appears that a one-off eye-opener about the metaphoric nature
of certain expressions is not sufficient to yield a long-term advantage in
retention. On the other hand, as might be expected, it appears that recurring
awareness-raising activities do have a long-tenn beneficial effect on memory
storage. One hypothesis is that the students whose metaphor awareness was
A second practical concern for language teachers may be the degree of moni-
toring that is required when students apply metaphor awareness to expand
their language resources. Iflearners have become aware of the prevalence
of a certain metaphoric theme in the target language, can they then use this
awareness to "generate" figurative expressions independently or should they
carefully confine their usage to expressions they have actually encountered in
the target language? So far, the proposal has been to use metaphor awareness
only in the latter sense, i.e., as a channeling device to comprehend, store, and
reproduce figurative language input.
The problem with having learners use an identified metaphoric theme to
generate figurative expressions is that the result may not be standard, con-
ventionallanguage. After all, it may be feasible to motivate idiomatic expres-
sions by relating them to a common metaphoric theme or source domain,
but it remains impossible to predict exactly what the idioms belonging to
that metaphoric theme will look like in a given language. In other words, it
is impossible to predict exactly how a particular language will instantiate a
given conceptual metaphor. If learners employed their recognition of that
metaphor to generate figurative expressions in the target language inde-
pendently, they would have to fully appreciate the chances of producing
"marked" language. For example, it does not follow from the well-docu-
mented existence of the metaphor ANGER IS HEAT in English that state-
ments like After all those insults, smoke came out of his ears and After this
morning s argument, she 50 been a barbecue all day are standard English
phrases, even though they may be quite comprehensible.
An additional argument in favour of confining a learner's output to the
figurative expressions s/he has actually encountered is the risk ofLI inter-
ference. When two languages are closely related, they will probably share
many metaphoric themes. On the one hand, recognising such shared
metaphoric themes may be helpful to the learner, because transfer from Ll
to the target language can then speed up the learning process. On the other
hand, transfer inevitably involves the risk of erroneous direct translations
(e.g., Cornell 1999; Swan 1997), and this risk appears especially high in lan-
guage learners who actually perceive the two languages as close (Kellerman
1987). Even "siblings" of idioms vary across languages. The Dutch and
French equivalents of the English idiom biting someone s head oj]; for
example, are iemand zijn neus afbijten and se faire bouJJer le nez, which both
depict noses rather than heads being bitten off. Knowledge of the existing
metaphoric themes of the target language does not entail mastery of its stan-
dard linguistic instantiations.
On the other hand, figurative language is (fortunately) not made up exclu-
sively of conventional expressions, and the above arguments in favour of
carefully monitoring students' language output are perhaps rooted in language
learning programmes that aim at native-like linguistic accuracy. It might be
worthwhile to investigate what would happen if learners were actually
encouraged to "take risks" and to exploit their knowledge of a prevalent
metaphoric theme in the target culture to "coin" figurative expressions
through creative thinking. Such an experiment would inevitably result in a
fair number of unconventional expressions, some of which would undoubt-
edly be considered by native speakers as wrong. Others, however, would
possibly be considered merely as original or perhaps even as poetic. A
research question here is to what extent native speakers comprehend and
tolerate such unconventional instantiations of a familiar metaphoric theme.
In a pilot experiment, English teachers at the University of Birmingham
were asked to rate expressions as (i) wrong English; (ii) deviant, but accept-
able English; or (iii) correct English. The expressions were presented in the
context of two fictitious e-mail messages:
Dear Bill,
I'm afraid I won't be able to join you on that hiking trip next weekend. I got
into trouble with Paula and I think I'd better spend the weekend with her to
make up, you see. The thing is, I forgot her birthday last Monday, which was
stupid because you know what a ferocious temper she has. To make matters
even worse, I went out for a couple of beers after work and I got home rather
late that night. So she wasn't exactly in a good mood when I fmally turned
up - without a birthday present. In fact,
Dear Bill,
I'm afraid I won't be able to join you on that hiking trip next weekend. I got
into trouble with Paula and I think I'd better spend the weekend with her to
make up, you see. The thing is, I forgot her birthday last Monday, which was
stupid because you know what an inflammatory temper she has. To make
matters even worse, I went out for a couple of beers after work and I got
home rather late that night. So she wasn't exactly in a good mood when I
finally turned up - without a birthday present. In fact,
Two of the options were existing English idioms (unleashing one;" anger
and breathing fire), two were close to these existing idioms (unchaining
one's anger and breathing flames) and two were "completely novel" cre-
ations (letting one 's anger out olits cage and being aflame-thrower). Please
notice that in each e-mail message the three options instantiated a
metaphoric theme that had been introduced earlier on through the phrases
ferocious temper and inflammatory temper, respectively.
The informants' responses to the non-standard phrases varied consider-
ably, as shown in table 1.
Breathing flames 2 7
Being a flame-thrower 7 2
(2 responses). Letting anger out ~r its cage was judged acceptable or correct
by no fewer than six respondents. The difference in appreciation may be
due to the fact that the scene of letting an animal out of its cage is a more
salient source domain experience for most people than a scene involving a
flame-thrower. Put simply, the flame-thrower phrase may have come across
as too far-fetched. Nevertheless, these preliminary data seem to suggest that
completely novel instantiations of an established metaphoric theme may
more easily be considered by native speakers as correct (but possibly
"poetic") language than slightly deviant versions of standard idioms.
Considering the fair degree of tolerance on the part of native speakers
attested in the above pilot experiment, it seems that it may be feasible to take
metaphor awareness beyond the realm of receptive vocabulary learning, after
all. Language learners might be aided by their understanding of a prevalent
metaphoric theme in the target culture to actively communicate abstract
notions through creative figurative language. The success of such an
endeavour would clearly depend on the learners' awareness of (i) the proba-
bility of cross-linguistic variation in conventionalised instantiations of the
given metaphoric theme and (ii) the possibility of cross-cultural variation in
the popularity of the metaphoric theme itself. After all, even when a meta-
phoric theme is shared by two linguistic cultures, it may be more popular in
one culture than the other (e.g., Boers and Demecheleer 1997; Boers 1999;
Emanatian 1995).
The decision whether metaphor awareness should serve only as a recep-
tive channel for vocabulary retention or whether it could also be used to
inspire students' language generation will ultimately depend on the objectives
of the language learning programme. If the programme aims principally at
native-like accuracy, then careful monitoring oflearners' figurative language
output will obviously be required. On the other hand, if the programme is
oriented towards using the language as a medium for intercultural commu-
nication (e.g., Byram 1997), then learners could possibly exploit their
knowledge of attested metaphoric themes to create figurative language
themselves. Moreover, overextension of a learned metaphoric theme may
still be rectified as the learner progresses and becomes more sceptical of L 1
borrowing (see below).
A third practical concern for language teachers may be what level of profi-
ciency their students need in order to successfully apply metaphor aware-
ness to vocabulary learning. For teachers who still consider metaphor to be
an ornamental device associated with poetry and literature, the use of meta-
phor in the classroom will probably appear appropriate only when teaching
advanced students. However, when metaphor is recognised as a basic cogni-
tive process which is at the root of a plethora of linguistic phenomena (poly-
semy, idioms, proverbs, etc.), then it may very well be relevant to all levels
of proficiency. Thinking and communicating abstract thoughts requires
metaphor. Although certain teaching methods, such as Total Physical
Response (Asher 1977), do tend to postpone abstraction at elementary levels,
it is difficult to imagine a complete ban on all abstract thought in a learning
situation involving (near-)adults. Whenever genuine communication takes
place, abstract ideas may be expressed, and thus figurative language may be
needed. Moreover, metaphor provides learners with a tool to extend the
meaning of simple, concrete words to denote more complex, abstract con-
cepts for which they have not yet acquired the precise tenns.
The aforementioned experiments about metaphor awareness and vocab-
ulary retention were all carried out with the participation of intennediate
learners of English. On the one hand, these Dutch- and French-speaking
students may have been aided in their processing of the presented meta-
phors by the close cultural relationship with English. On the other hand,
they may also have been impeded in their processing of the presented figu-
rative language by gaps in their lexical knowledge. Further experimental
research is clearly needed to measure the effects of an enhanced metaphor
awareness in (i) elementary or advanced language learners and (ii) learners
of a language that is more distant from their own.
It is obvious that elementary language learners would face most difficulty
in applying a strategy of metaphor awareness to newly encountered expres-
sions, quite simply because they would often lack the lexical knowledge
needed to interpret such expressions in the first place. For instance, in order
to independently make sense of she was fuming, the learner would need to
understand the word fuming. In order to recognise the metaphoric nature of
many expressions, one needs knowledge of the literal (or prototypical)
senses of polysemous items. For example, an elementary learner who has
never encountered any spatial sense of the preposition beyond may not be
A fourth practical concern for language teachers may be whether the technique
of metaphor awareness is powerful enough to accommodate all individual
learners, even if they share the same level of proficiency. While the afore-
mentioned learning experiments about vocabulary retention through metaphor
awareness reveal beneficial effects in the experimental groups as a whole, they
also show considerable differences between the scores of individual learners.
F or example, in the experiment set up to measure learners' expansion of
vocabulary to describe upward and downward trends (soar, plunge, etc.), the
scores of individual participants ranged from 3 to 14 reproduced expressions.
Such variation calls for caution to never treat a group of learners as a homo-
geneous population. Individual differences are inevitable along multiple
affective and cognitive parameters. One parameter is that of cognitive styles.
A cognitive style can be roughly defined as an individual's consistent
strategy to organise and process information. A number of cognitive style
dimensions or continua have been identified in psychology (e.g., Schmeck
1988; Witkin and Goodenough 1981) and two of these appear to be of special
interest to our discussion: the analytic / holistic continuum and the verbaliser /
imager continuum (e.g., Riding and Cheema 1991).
The analytic / holistic dimension (e.g., Bever 1975) refers to an individ-
ual's preference for processing information either as separate parts or as
large integrated chunks. Analytic individuals tend to break down a problem
into distinct parts, whereas holistic individuals tend to maintain a more
global picture. The verbaliser / imager dimension refers to a subject's pref-
erence for thinking either in words or in pictures (e.g., Paivio and Harshman
1983). Both of these dimensions may also have a profound impact on the
way individuals process metaphors and figurative language.
In a dual experiment (described in detail in Boers and Littlemore 2000),
71 university students were asked to explain the existence of a couple of
well-established metaphoric themes. The respondents differed in their pre-
ferred strategies to explain the metaphors and these strategies correlated
with their cognitive styles as "diagnosed" by means of the Riding (1991)
computer-assisted test of cognitive styles. Very briefly, this test measures
analytic processing by presenting the participant with a simple shape and a
complex shape, placed side by side. The participant's task is to decide as
quickly as possible whether the simple shape is contained within the complex
shape. To measure holistic processing, the participant is asked to decide
whether two shapes are identical. Verbal and imagery processing are meas-
ured by presenting pairs of words which the participant has to assess
whether these are the same type or the same colour.
Firstly, participants who were found to have an analytic cognitive style
appeared best at tracing a given metaphor back to a distinct, concrete source
domain. They clearly distinguished literal from figurative usage. Holistic
respondents, on the other hand, appeared to find it harder to identify a dis-
tinct source domain at the origin of a given metaphor. In their explanations
they were more likely to "blend" their conception of different domains
without making a clear distinction between literal and figurative.
Secondly, imagers were more likely than verbalisers to explain the given
metaphors by referring to concrete scenes. For example, they explained the
metaphoric theme ECONOMIC COMPETITION IS RACING (lagging behind;
the race for market share; etc.) by calling up the scene of a busy businessman
who is in a hurry to get to prospective clients or to new markets before his
competitors do. Likewise, imagers typically explained the ECONOMIES ARE
MACHINES metaphor (exchange-rate mechanisms; monetary tools; etc.) by
calling up the scene of a Fordist assembly line with machinery.
Although the aims of this experiment were not directly related to FLA,
the observed variation in individuals' processing of metaphors does suggest
that language learners' are also likely to have different approaches to unfa-
miliar figurative expressions in the target language. Due to their cognitive
style characteristics, some learners may be more susceptible to the strategy
of metaphor awareness than others. The findings lead us to hypothesise that
an enhanced metaphor awareness will probably be most beneficial to learners
with an analytic and imager cognitive style. Analytics are probably most
capable of recognising the metaphoric nature of an expression or the figura-
tive usage of a polysemous word by comparing it with a distinct source
domain or literal usage. The identification of distinct source domains (or
metaphoric themes) behind sets of expressions provides a framework of
vocabulary organisation, which is known to facilitate memory storage.
Imagers are probably most capable of associating a novel figurative expres-
sion with a mental picture or concrete scene, and imaging and concreteness
have also been shown to facilitate retention.
Again, it is clear that a lot of experimental research would be required to
test these hypotheses. One major complication in such research is the likeli-
hood of gradual shifts in individuals' learning styles. After all, a learning
process is supposed to have an effect on the learner's cognitive processing.
Furthermore, even if future research turns up more conclusive evidence of
cognitive style differences in learners' susceptibility to the strategy of en-
hanced metaphor awareness, then many questions will remain as to possible
implications for educational practice. It goes without saying that a language
course should ideally cater to all individual learning styles present in a given
target group. In practice, this means that different approaches to learning
can complement one another.
A fifth practical concern for language teachers may be that not all idioms
lend themselves equally well to explicit imaging techniques. While the
the figurative usage of hurdle in the government is facing major hurdles can
be clarified by referring to the literal sense of the word (e.g., a steeplechase
context).
A complementary method to motivate an idiom is to retrace its likely
etymological origin (e.g., Boers 2001). For example, it is feasible for teach-
ers in a classroom context to quickly clarify the use of under in under one 50
own steam by referring to the era of steam propulsion on ships and trains.
Still, we have to acknowledge the existence of idioms whose origins have
become too obscure to be of any use to the average learner. For instance,
tracing the origin of under way hardly seems worth the effort in general
EFL classrooms.
Further empirical research is clearly required if we want to help teachers
(and learners) decide whether a given idiom is a suitable candidate for
imagery processing. A research team at the University of Antwerp and the
Erasmus College of Brussels has recently embarked on just such a project.
As part of a computer-assisted vocabulary acquisition programme, large
numbers of idioms are processed by students along alternative cognitive
channels, including a cognitive channel that exploits mental imagery. By
assessing the students' ease at using a particular channel for each idiom, and
by assessing the students' rate of retention of each idiom after using a par-
ticular cognitive channel, the study hopes to generate a practical repertoire
of figurative idioms for which imagery processing can be recommended on
the basis of empirical evidence. At a more theoretical level, the collected
data may allow us to identify more precisely what features of idioms con-
tribute to "imageability" in the first place.
A final concern (closely related to the previous one) may be the problem of
categorising figurative expressions under precise metaphoric themes. After
all, it has been argued that one of the advantages of an enhanced metaphor
awareness is the mental framework it provides for organising figurative lan-
guage input. Categorising idioms under their common metaphoric themes
lends structure to the influx of figurative language which may at first sight
seem completely arbitrary and unsystematic.
Unfortunately, different conceptual metaphors inevitably intersect, so
that different ones may be at play simultaneously. For example, an expres-
sion such as slimmingjlabby workforces may be presented by teachers as an
instantiation of HEALTH, FITNESS, and SPORTS metaphors. Moreover, rather
specific metaphoric themes can be subsumed under more generic ones. For
example, an expression like We are catching up with our competitors can be
presented as an instantiation of a RACING metaphor as well as a more general
SPORTS metaphor. The identification of conceptual metaphors is an issue of
ongoing debate in cognitive semantics (e.g., Grady 1997, 1999) and an in-
depth discussion is clearly beyond the scope of this applied linguistics article.
The relevance for SLA, though, is that recurring difficulties in matching
figurative expressions to precise and unique metaphoric themes could result
in confusion and scepticism on the part of the learner. The advantage of
establishing some kind of order in the complex world of figurative language
could be lost. To avoid this, it seems advisable (in initial stages) to draw
learners' attention principally to "clear" cases, i.e., expressions whose source
domain can be pinpointed pretty unambiguously.
Unfortunately, for numerous conventional figurative expressions the link
with the original source domain has become very vague. For example, few
language users are likely to recognise economic recovery as being derived
from the domain of health, unless the context points to this source domain
(e.g., After a long period of illness the economy is recovering at last).
Checklists for the recognition and categorisation of figurative expressions
in real discourse are being developed (e.g., Steen 1999a, 1999b), but to my
knowledge these have not yet been geared to FLA needs.
Apart from serving as a framework of organising figurative language
input, the strategy of metaphor awareness serves to enhance imaging and con-
creteness. As has been mentioned above, metaphoric themes (or conceptual
metaphors) vary in their degrees of abstraction: abstract, generic metaphors
subsume more concrete, specific ones. While the former contain "bare" image
schemas (PATH, CONTAINER, etc.) (e.g., Lakoff 1990), the latter contain
more elaborate or "richer" source domain information. In order to facilitate
imaging in an SLA context, it seems advisable to give preference to the
more specific metaphoric themes over the more generic ones. For example,
an expression such as Go on with your work, which reflects the PATH
schema contained in the generic ABSTRACT ACTIVITY IS MOTION
metaphor, is not likely to evoke rich imagery. On the other hand, expres-
sions such as The enterprise has been blown off course, The enterprise has
been derailed, and The enterprise hasjaken off, which reflect more specific
motion metaphors involving ships, trains, and airplanes, appear much better
candidates for explicit metaphor processing.
6. Conclusions
In this chapter I have tried to establish the conditions under which an en-
hanced metaphor awareness is likely to be beneficial to language learners'
vocabulary expansion. On the way, we have briefly visited six unexplored
areas of investigation to gauge the following research questions.
References
Asher, James
1977 Learning Another language through Actions: The Complete Teacher s
Guidebook. Los Gatos, CA: Sky Oaks Productions.
Baddeley, A.
1990 Human Memory: Theory and Practice. Needham Heights, MA:
Allyn and Bacon.
Bever, T. G.
1975 Cerebral asymmetries in humans are due to the differentiation of two
incompatible processes: holistic and analytic. Developmental Psycho-
linguistics: Communication Disorders 26(3): 251-262.
Boers, Frank
1996 Spatial Prepositions and Metaphor: a Cognitive Semantic Journey
along the Up-Down and the Front-Back Dimensions. (Language in
Performance 12.) Ttibingen: Gunter Narr Verlag.
1999 When a bodily source domain becomes prominent: the joy of counting
metaphors in the socio-economic domain. In: Raymond W. Gibbs
and Gerard J. Steen (eds.), Metaphor in Cognitive Linguistics, 47-56.
(Current Issues in Linguistic Theory 175.) AmsterdamlPhiladelphia:
John Benjamins.
2000a Metaphor awareness and vocabulary retention. Applied Linguistics
21(4): 553-571.
2000b Enhancing metaphoric awareness in specialised Reading. English
For Specific Purposes 19(2): 137-147.
2001 Remembering figurative idioms by hypothesising about their origins.
Prospect 16(3): 35-43.
Boers, Frank and Murielle Demecheleer
1997 A few metaphorical models in (western) economic discourse. In:
Wolf Andreas Liebert, Gisela Redeker and Linda Waugh (eds.),
Discourse and Perspective in Cognitive Linguistics, 115-129.
(Current Issues in Linguistic Theory 151.) Amsterdam! Philadelphia:
John Benjamins.
Boers, Frank and Murielle Demecheleer
1998 A cognitive semantic approach to teaching prepositions. English
Language Teaching Journal 52(3): 197-203.
2001 Measuring the impact of cross-cultural differences on learners' com-
prehension of imageable idioms. English Language Teaching Journal
55(3): 255-262.
Boers Frank and Jeannette Littlemore
2000 Cognitive style variables in participants' explanations of conceptual
metaphors. Metaphor and Symbol 15(3): 177-187.
Byram, Michael
1997 Teaching and Assessing Intercultural Communicative Competence.
Clevedon: Multilingual Matters.
Johnson, Mark
1987 The Body in the Mind: The Bodily Basis ofMeaning, Imagination and
Reason. ChicagolLondon: University of Chicago Press.
Kellerman, Eric
1978 Giving learners a break: native language intuitions as a source ofpre-
dictions about transferability. Working Papers in Bilingualism 15:
59-92.
1987 Aspects of transferability in second language acquisition.
Unpublished PhD. Dissertation, University ofNijmegen.
K6vecses, Zoltan
1986 Metaphors of Anger; Pride and Love: a Lexical Approach to the
Study of Concepts. AmsterdamlPhiladelphia: John Benjamins.
K6vecses, Zoltan and Peter Szabo
1996 Idioms: a view from Cognitive Semantics. Applied Linguistics 17(3):
326-355.
Lakoff, George
1987 Women, Fire and Dangerous Things: What Categories Reveal about
the Mind. ChicagolLondon: University of Chicago Press.
1990 The Invariance Hypothesis: is abstract reason based on image-
schemas? Cognitive Linguistics 1(1): 39-74.
Lakoff, George and Mark Johnson
1980 Metaphors we Live by. Chicago/London: University of Chicago
Press.
Lazar, Gillian
1996 Using figurative language to expand students' vocabulary. English
Language Teaching Journal 50(1): 43-51.
Lennon, Paul
1998 Approaches to the teaching of idiomatic language. IraI36(1): 12-30.
Lindner, Susan
1981 A lexico-semantic analysis of verb-particle constructions with Up
and Out. Unpublished PhD dissertation, University of California San
Diego.
Lindstromberg, Seth
1997 English Prepositions Explained. Amsterdam/Philadelphia: John
Benjarnins.
Low, Graham
1988 On teaching metaphor. Applied Linguistics 9(2): 125-147.
Murphy, G.
1996. On metaphoric representation. Cognition 60: 173-204.
Paivio, A. and R. Harshman
1983 Factor analysis of a questionnaire on imagery and verbal habits and
skills. Canadian Journal ofPsychology 33 (1): 17-28.
Riding, R. J.
1991 Cognitive Styles AnalYSis. Birmingham: Learning and Training
Technology.
Szilvia Csabi
1. Introduction
The present paper deals with the method of vocabulary teaching, using the
principles developed in cognitive linguistics. The focus is on polysemy and its
implications for teaching English as a second language. Cognitive linguistic
research on polysemy (i.e., "the association of two or more related senses
with a single linguistic form" [Taylor 1989: 99]) has suggested that the
meaning structure of polysemous words is motivated and can be accounted
for in a systematic way. This insight can be employed in applied linguistics
and can be put to use in teaching as well as learning foreign languages.
The claim proposed here is that teachers' and learners' awareness of the
cognitive mechanisms making up the network of senses is useful in the lan-
guage learning process. It is not claimed here that traditional vocabulary
learning processes, such as memorization, should be replaced. Rather, the
point of the study is to support the idea that, besides memorization, awareness
and acquisition of the cognitive structure of word meanings aids teaching
and learning.
Words with several senses often cause problems for teachers and learners
of English since they are often seen as unrelated and unsystematic. This belief
probably discourages teachers and students to deal with polysemous words
in the classroom. The strategy commonly employed by teachers is to not
deal with the various senses of a certain word all at once, but to explain the
specific senses when they turn up. However, a more fruitful strategy, result-
ing in the easier mastery of the target language, would be to provide the
explanations and motivations for the related senses of given words together.
In order to support the appropriateness of the present approach to vocab-
ulary teaching and learning, two experiments are reported on examining
whether the explicit knowledge of conceptual metaphors, metonymies, and
conventional knowledge present as motivational factors in the target lan-
guage facilitate learning word meanings in the target language. Classroom
research on teaching polysemous words such as hold and keep and idioms in
which they appear to secondary school students appears to support the appli-
cation of cognitive linguistic principles in language teaching and learning.
In sum, this paper claims that learners who know how certain conceptual
mechanisms (such as conceptual metaphors and metonymies) structure the
meanings of certain polysemous words and idioms will acquire their mean-
ings more easily, compared to other learners who are not familiar with these
mechanisms. On these grounds, it is suggested that the motivations of word
meanings should be clarified to students in the language classroom to
achieve better results.
guistics in the teaching and learning idioms, more specifically, phrasal verbs
with up and down. Their findings support the idea that it is advantageous to
use the cognitive paradigm in the teaching and learning of idioms in L2.
The use of the cognitive paradigm in teaching and learning builds upon
the principle of motivation, which is a central phenomenon in cognitive lin-
guistics. Lakoff (1987) defines motivation as follows: "The relationship
between A and B is motivated just in case there is an independently existing
link, L, such that A-L-B 'fit together'. L makes sense of the relationship
between A and B" (448, italics in the original). The reason for its centrality is
that "[ijt is easier to learn something that is motivated than something that
is arbitrwy. It is also easier to remember and use motivated knowledge than
arbitrary knowledge" (Lakoff 1987: 346, italics in the original). Possible
motivations for word meanings can be provided by conceptual metaphor
(CONCEPT A IS CONCEPT B), conceptual metonymy (CONCEPT A FOR CON-
CEPT B), conventional knowledge, image schemas (e.g., UP-DOWN schema),
and construals (Lakoffand Johnson 1980; Lakoff 1987; Langacker 1987).
Motivation plays an important role in a number of phenomena in cognition
and language. Lakoff observes that polysemous words and most idioms arc
motivated, since their meanings make use of and are consistent with certain
already existing patterns (1987: 438). Naturally, not all word meanings and
idioms are motivated, but in the majority of cases we can find a link between
form and meaning. As Lakoff (1987: 316) argues,
Polysemy occurs when a single word has more than one meaning - and when
those meanings are systematically related. Systematic relationship is crucial
here. The two meanings of bank - place where you put your money and the
edge of a river - are not systematically related. Such cases are called hom-
onyms. Cases of poly se my are cases like warm, which refers to both to tem-
perature and to clothing that makes you feel warm. Another example would
be newspaper, which can name either what you read at the breakfast table or
the company that produces it.
for the learners of English since they can easily be confused. They appear to
be used in a variety of senses, which are probably difficult to acquire
because they are seemingly unrelated. In addition, learners are usually only
taught their most frequent meanings (as in She was holding an umbrella or
You can keep the change). However, gaps in knowledge of further senses
may often turn out to be impediments to fluent interaction with and clear
understanding of the speech of native speakers.
3. Method
The study designed to investigate the effects of learners' (lack ot) knowl-
edge of sense motivations consisted of two experiments. The experiments
employed the same materials and similar, though not identical, procedures.
First, a pilot study was conducted in the fall of 1997 in a Budapest secondary
school (the same school where the later experiments were carried out). The
subjects of the pilot study were 1ph-graders (age 17) who were on an ad-
vanced level of English, and who had four English lessons a week. The aim
of the pilot study was to see whether the experiment is conductible, and if
so, then which test items can be used in the redesigned experiments. After
the pilot study, the materials and the procedures were changed and their
final version was used in the experiments reported below.
3.1. Subjects
The classes selected for the study performed in the winter of 1998 included
fifty-two students enrolled in a Budapest eight-class secondary school. Half
of them were in their 8tll grade (ages 13-14) and half of them were in their 9th
grade (ages 14-15). The students belonged to four learner groups (in what
follows: Group A, B, C, and D) with thirteen students in each at the time of
the actual research. All the students involved in the study were native
speakers of Hungarian and learned English as a second language. Groups A
and B - the groups in the 8th grade - were of roughly equal proficiency
(intermediate level); they had been studying English for two and a half
years, and they had 4-5 English lessons a week. Groups C and D - the
groups in the 9th grade - were also of roughly the same level of proficiency
in English (intermediate level). The members of these two groups had been
studying English for three and a half years and had 4-5 lessons a week.
Group A was the experimental group and Group B was the control group in
the first experiment. Group C was the experimental group and Group 0 was
the control group in the second experiment.
3.2. Procedures
3.2.1. Instruments
The instrument used in the research was a test containing twenty-two items
related to hold and keep (see Test in the Appendix). Tt focused on (a) hold
and keep as polysemous items; (b) phrasal verbs including hold and keep;
(c) idioms containing hold and keep. Naturally, students in every group
were familiar with the most frequently used meanings of hold and keep, as
illustrated in He was holding a kn~fe in one hand and a fork in the other, She
held her daughter s hand as they crossed the road, I held the baby in my
arms or in Here s a jive-pound note - you can keep the change and I keep
all her letters. That is, the central meanings of hold and keep were already
known to students - nevertheless, they were also included in the teaching and
testing process since further, figurative meanings and idiomatic meanings
largely depend upon these meanings. The information provided by the teach-
ers indicated that all the other items in the test were unknown to the students.
The senses of hold and keep and the idioms in which they occur were all
pretaught to each group. The language used during the explanations was
Hungarian for practical reasons. Since the researcher had no previous teach-
ing experience with the groups involved in the study, it would have created
an uncontrollable factor to use English words in the explanations that were
not known to the students.
The general teaching procedure and the administration of the test were
divided into three parts in each group: First, some senses of hold and keep
were presented to the groups. Students were instructed to memorize the
words and their meanings. Following this, they were required to complete
the first ten sentences of the test. Either hold or keep had to be placed in the
gaps. Secondly, phrasal verbs containing hold and keep were presented to
the classes, whose meanings students had to memorize. Afterwards, the sec-
ond part of the test had to be completed, in which the phrasal verbs had to
be inserted. Thirdly, multiword idioms containing hold and keep were pre-
sented to students, who memorized their meanings. After this activity they
had to fill in the third part of the test, which contained more than one sen-
tence describing a situation appropriate for the use of a given idiom. The
items to be inserted in the gaps were not written on the test sheet, so students
had to retrieve them from memory. Shortly after the teaching and testing
procedure described above, a post-test was administered to each group, that
is, students were required to complete the same test as they had to fill in on
the day of the treatment.
The only difference between the experimental and the control groups
was the presentation of the material to be taught. Groups A and C, the experi-
mental groups, were taught in accordance with the cognitive linguistic
approach to meaning described in detail below. This means that instead of
teaching them the English expressions with their Hungarian equivalents,
only the most important motivating factors were explained to them, and they
memorized the items on the basis of their knowledge of sense motivations.
In contrast to this, Groups Band D, the control groups, were only taught the
English words and their Hungarian equivalents, and they relied only on this
information when memorizing the expressions and their meanings.
The theoretical background for the systematic explanation of the sense
motivations of hold and keep was provided by Lexical Network Theory
(Norvig and Lakoff 1987). This theory sees "a lexical item as a network of
minimally differing senses, with links of a small number of types" (Norvig
and Lakoff 1987: 195). Hold and keep have several senses that are related to
each other, and no abstract, general meaning can be discovered which covers
all possible senses. A plausible explanation for the systematicity in the dif-
ferent senses of both hold and keep may be given in terms of a network
"such that each sense is a minimal variant of some other sense" (Norvig and
Lakoff 1987: 197, italics in the original). This means that if sense A is a min-
imal variant of sense B, there is only a single significant difference between
them from which their other differences can be predicted. According to
Norvig and Lakoff(1987: 197-198), the following links exist which connect
the various senses of polysemous words: image-schema transformation
links, metaphoric links, metonymic links, frame-addition links, semantic
role differentiation links, and profile shifts.l
Lexical network analysis provides a fine-grained structure of word mean-
ing and lends itself to the graphical representations of meaning structures.
Experiments concerning prepositional meanings reported by Sandra and
Rice (1995) seem to prove the legitimacy of lexical network analysis since
statistically significant results suggest that "language users make rather fine-
grained distinctions, much in the way that prepositional network modellers
do" (124). The empirical evidence obtained from the experiments suggests
that language users are able to perceive certain usage types as homonymous
and others as related to each other, similar to the way proposed by linguists
(Sandra and Rice 1995). This supports the validity of the analysis.
Besides the force dynamic pattern applied in the physical sense of hold up
(Ralph held up his hand), the metaphorical sense of hold up is motivated by
the metaphor OBSTRUCTION IS UP. Thus, we can use hold up in The whole
thing was held up about half an hour and The criminals held up the train.
F or the purposes of the experiment, the following idioms were chosen:
hold one 50 tongue, hold one 50 temper, and hold one 50 head up. Hold one 50
tongue implies force dynamic patterns (the tongue would move, but it is not
allowed to move) as well as control through the metaphor CONTROL IS
HOLDING SOMETHING IN THE HAND. We also have a conventional image
of a person holding his or her own tongue. Evidently, tongue metonymically
stands for speaking. Hold one 50 temper is similar in some respects to hold
one 50 tongue since it also indicates force dynamic patterns and control, due
to the fact that emotions are often understood as forces that have to be con-
trolled (Kovecses 2000). Finally, hold one s head up receives motivation
from the emotion metaphor THE CONCEITED PERSON IS UP/HIGH (Kove-
cses 1990). This is again an easily imaginable idiom, with a conventional
image in which a person's head is held high up in the air.
The meanings of the above idioms appear to be motivated primarily by
metaphorical extension. The senses of hold render it possible that a number
of idiomatic expressions develop whose meanings are also motivated, in
part by the meaning structure of hold.
Further senses of keep are linked primarily by profile shift, that is, something
which is backgrounded in one sense is fore grounded in another, minimally
related sense (Norvig and Lakoff 1987: 190). In the minimally different
sense keep-2 explicated by She kept a diary for over 20 years, continuity,
the durative component, is foregrounded. In keep-3, illustrated in She kept
her promise / her word / the secret both the force dynamic pattern and the
durative component are profiled (e.g., the person does not let anyone else
know the secret). Keep-4, a further variant of keep-I, profiles not only the
aspect of continuity but also the aspect of temporary possession and main-
taining a state as well as in It costs more each year to keep a house, He
scarcely earns enough to keep himself and his family.
Idioms with keep such as keep in and keep out (of); keep one 50 fingers
crossed, keep somebody at arm 50 length, keep something under one 50 hat
profile the force dynamic pattern. The meanings of the phrasal verbs keep in
and keep out relate to the meaning of keep in which the durative and the
force dynamic components are profiled. The particles in and out evoke the
image of a container, a bounded area, with things that can be either in or out
of it. On the one hand, keep in can metaphorically refer to emotions as in
He could scarcely keep in his indignation since human bodies are often
understood as containers for emotion, and therefore the CONTAINER meta-
phor is often utilized in talking and thinking about emotions (K6vecses
2000). On the other hand, keep in can also refer to a culturally motivated
situation in school illustrated in She was kept in for an hour for talking in
class. In this case, the building of the school, more specifically the class-
room itself, is the metaphorical container with the pupil inside it. In line
with the above, keep out focuses on the outside of the container, that is,
what is out of the container. Usually there is a boundary that the relevant
entity should not cross, as in The sign said "Ministry ofDeji:mse - Danger -
Keep out!" Keep that dog out of my study! and That child seems incapable
of keeping out of mischief
Keep one 50 fingers crossed, keep somebody at arm 50 length, and keep
something under one 50 hat also focus on force tendencies and continuity
implied by keep. The aim of the agent is to maintain the state or position of
the fingers, somebody, or something, respectively. Keep one 50 fingers
crossed implies the image of having the fingers in a special position for a
long time, which stands metonymically for being anxious or worried about
somebody. The idiom keep somebody at arm 50 length has metaphorical
motivation as well. Lakoff (1987: 447) argues that a conventional image
(the arm, which is chest high, is oriented forward with respect to the body;
the hand is open and the palm is facing away from the subject; the angle of
the hand relative to the forearm is roughly 90 to 135 degrees; the ann muscles
are tense; and the person being kept at arm's length is facing the subject), and
two metaphors (INTIMACY IS PHYSICAL CLOSENESS and SOCIAL/PSYCHO-
LOGICAL HARM IS PHYSICAL HARM) motivate the meaning of this idiom
("keep someone from becoming intimate, so as to protect oneself from social
or psychological harm"). Keep something under one s hat also implies the
existence of a conventional image with a person and his or her hat, under
which there is something hidden. This image is complemented by the
metaphor KNOWING IS SEEING. The rationale is that if something cannot be
seen, it cannot be known either - this is why the secret is hidden under the
person's hat. This is the logic that two other idioms with keep (keep the lid
on and keep something under wraps) also utilize.
3.3. Experiment 1
a circle with an X inside and outside indicated the phrasal verbs keep in/out).
After the explanation part, students had to memorize the expressions and
their meanings, and they had to complete the six sentences given with the
phrasal verbs they had just learned. The procedure took ten minutes.
In the last part of the lesson, idiomatic expressions such as hold one S
tongue or keep something under one s hat were presented to the class. The
expressions and their schematic, but representative, drawings were sketched
on the blackboard. These illustrated the mental images inferred from the
literal meanings of the idioms (thus, the meaning of keep something under
one s hat was indicated by a hat and something under it). The meanings
were explicated and memorized, and afterwards the students were instructed
to complete the last task of the test.
During the explanation parts, body language was also used to strengthen
the students' conception of meanings. The illustrative part of the explana-
tions (the use of drawings in addition to body language) appeared to be very
useful for the students' better understanding of meanings. In several cases,
students themselves found out the exact meanings of the expressions in
question.
In the control group (Group B), the lesson also consisted of three pat1s.
The difference in the procedure was that the motivation for the meanings
was not given. Instead, the words and expressions were written on the black-
board with their Hungarian equivalents. In the first part of the lesson, the
selected meanings were put on the blackboard in Hungarian. They were clar-
ified in a few words, but nothing was said about the way the meanings are
connected. Students had to memorize the words and their senses, and after-
wards they had to complete the first ten sentences of the test. This took fifteen
minutes altogether. 3 Afterwards the phrasal verbs were written on the black-
board with their Hungarian equivalents. Again, the meanings were clarified,
but motivations were not given. After the explanation and the memorization
procedure, students were instructed to complete the second gap-filling exer-
cise. Having spent ten minutes on this task, idioms and their Hungarian
equivalents were presented and written on the blackboard, which the stu-
dents had to memorize. Finally, they were asked to fill in the last part of the
test using the idioms learnt just before. This last task lasted fifteen minutes.
One day later the test was re-administered to the two groups - naturally,
it was not known to them that they would have to work with the learned
material again. They were instructed to fill in the tests in ten minutes with-
out prior mention or recall of the given expressions.
3.4. Experiment 2
4. Results
The tests and the post-tests of all groups were analyzed, and the number of
correct answers was counted. Expressions in the second and third parts of
the test which were not entirely correct (e.g., keep this under hat or keep fin-
ger) but in which keywords (hold and keep) were given correctly were
counted as correct answers, although with indication that these answers
were not entirely correct. The reason for not excluding these answers was to
see whether students were able to differentiate between the senses of hold
and keep. The percentage values for the total number of correct answers on
the tests written by Groups A, B, C, and D are given in Table 1 and 2 in the
Appendix. The percentage values of the distribution of entirely correct and
not entirely correct (keyword-type) answers are supplied in Table 3.
The data from the two pairs of groups were then statistically compared
separately for each pair of tests. The tests of Group A were compared to the
tests of Group B (the first test of Group A to the first test of Group B, and the
post-test of Group A to the post-test of Group B). The same procedure was
applied in the case of Groups C and D. Not all the assumptions of the t-test
were met in the research; therefore the Wilcoxon Rank Sums test, a non-
parametric test, was chosen to make comparisons between Groups A and B,
as well as C and D. The null hypothesis was the starting point, according to
which there was no significant difference in the ranks assigned to the per-
formance of the groups, Groups A and B on the one hand, and Groups C and
D on the other hand, in any pair of tests.
The nondirectional Rank Sums test indicates that the differences between
the tests of Groups A and B as well as those between the post-tests of Groups
A and B are significant. The differences between the tests and between the
post-tests of Groups C and D are also statistically significant. In the test
written by Groups A and B the a level is at a < .01, and in the post-test of
these two groups the a level is at a < .01 again. In the test completed by
Groups C and D the a level is at a < .02, and in the post-test of these groups
the a level is at a < .01. The probability, then, is less than 2 % in general
that the differences between the group pairs occurred by chance alone.
Therefore, the null hypotheses of no difference can be rejected. Thus, the
statistically significant difference between the experimental group and the
control group in both experiments clearly shows that students in the experi-
mental groups outperformed those in the control groups.
In addition to being significant, the differences are meaningful because
they are consistently large in the case of both experiments. Ifwe take into
consideration the percentage values, the divergence between the two tests
taken by Group A (one on the day of the experiment and the other a day
later) shows that their performance decreased by 9.44%, whereas the per-
formance of the control group, Group B, decreased by almost the double of
this value, 18 %. The same tendency is observable in the case of Groups C
and D, where the difference of the performance of Group C was 8% after
two days, and where the difference of the control group performance
decreased by almost the double of this value after two days. The trend that
the post-test values are somewhat higher in the case of Groups C and D
(which they took two days after the first test) than in the case of Groups A
and B (which they took one day after the first test) can be attributed to the
fact that the separation of the material to be learned was beneficial to their
performance.
Table 3, which contains the percentage values of entirely correct and
keyword-type answers for each exercise of the test, shows that both experi-
mental groups did better in all three exercises than the matching control
groups as far as the total number of correct answers is concerned. In the sec-
ond exercise of the first test, both experimental groups gave good answers
in 100 % of the cases, and in the first exercise of the first test Group A wrote
good answers in 99.2% of the cases - which is very close to 100%. This
never happened in the case of the control groups. The highest value that
Group C achieved was 97.3% in exercise 2 of the first test, which is admit-
tedly a high value. The lowest values reached on any of the exercises are
attributable to the control groups: Group B achieved 52.5 % in exercise 3 on
the post-test (which is the lowest value of all), and Group D got only 68.4 %
of the correct answers in the first exercise on the post-test. Generally speak-
ing, the third exercise, where the multiword idioms had to be filled in,
appeared to be the most difficult exercise since the number of correct results
was the lowest there. Possibly the strings of words in the idioms were too
long to remember precisely since students only had a very short time to see,
understand, and memorize them. Still, the tendency was that the experimen-
tal group did better on this part of the test also - although it has to be noted
that group D, the control group, had a higher number of entirely correct
answers in exercise 3 than Group C, the experimental group; however, the
total number of good answers was still higher in the case of Group C.
5. Discussion
A further question that the study bears on is the effect of the strategy of
transfer from LI to L2. Hungarian has two words (jog and tart) that have
similar meanings. However, the meanings of the Hungarian words are not at
all identical with those of hold and keep. Rather, their meanings only partly
cover those of the English words. The network of meanings structuring fog
and tart appears to be different, and there are different metaphorical links
also. Thus, there are no one-to-one correspondences between the meanings of
word pairs. Also, the motivations for the English expressions do not always
and necessarily match the motivations for their Hungarian equivalents.
Nevertheless, the fact that the meanings of the selected idioms may make
use of similar conceptual motivation in English and Hungarian does not
ensure that their linguistic forms will be exactly the same. Thus, the idioms
do not necessarily correspond word for word in English and Hungarian.
Consequently, students could not simply translate the Hungarian expres-
sions into English.
Thus, students' awareness of conceptual links is beneficial in compre-
hending and using the target language adequately. This is in line with
Carter's suggestion (1993: 148) that "learning a language involves under-
standing something of that language [which] has to be explicitly taught".
Carter also claims that language awareness aids the development of inferen-
tial, interpretative, and analytic skills. These claims also support the view
taken by Widdowson (1990) according to which it is probable that "a con-
scious awareness of how language works and the subjection of their experi-
ence to analysis would suit [learners'] cognitive style, increase motivation by
giving added point to their activities, and so enhance learning". Widdowson
also adds that the skill of comparing L 1 with L2 would engage students in
rational inquiry advocated in the curriculum. Thus, a possible answer to
these needs of language teaching may be provided by applying the cognitive
approach to language, and within that, vocabulary teaching.
6. Conclusion
This paper aimed at showing that cognitive linguistics can offer language
teachers and learners a helpful approach to deal with polysemous words and
idioms. This approach can explicate motivations of senses and systemati-
cally link them. Some meanings of the polysemous words hold and keep and
some idioms used in the experiments were shown to be motivated within
the framework oflexical network analysis. In order to see whether language
learning is more fruitful and permanent if it operates with conceptual mech-
Acknowledgements
Appendix
Test
Tables
Table 1. Percentage values for the total number of correct answers (Groups A & B)
Group A GroupB
Test Post-test Test Post-test
96.15 % 86.71 % 88.1 % 70.6%
Table 2. Percentage values for the total number of correct answers (Groups C & D)
GroupC GroupD
Test Post-test Test Post-test
95.71 % 87.7% 88% 71.2%
Table 3. Percentage values for the number of entirely correct answers (ECA) and
keyword-type answers (KTA)
Notes
1. Description of the links used in the analysis (cf. Norvig and Lakoff 1987):
- Metaphoric links (i.e., links established by metaphoric mappings that exist
independently of the given lexical item), e.g., The baby took the toy jrom its
mother and He took a glance at the book: PERCEIVING IS RECEIVING metaphor.
- Metonymic links (i.e., links established by metonymic mappings that exist
independently of the given lexical item), e.g., The love between them is strong:
LOVE FOR THE RELATIONSHIP IT PRODUCES metonymy.
- Profile shift (i.e., instances where what is backgrounded in one sense is fore-
grounded in a minimally related sense), e.g., 1 took a punch at him and 1 took a
punch from him: result profiled in the second example.
2. It is outside the scope of this study to consider each meaning of hold and keep-
rather, those meanings that are made use of in the experiments are focused on.
3. The shorter time spent on this task is due to the fact that the explanation proce-
dure was shorter and less detailed.
References
Carter, Ronald
1993 Language awareness and language learning. In: M. Hoey (ed.), Data,
description, discourse, 139-150. London: HarperCollins Publishers.
Deignan, Alice, Danuta Gabrys, & Agnieszka Solska
1997 Teaching English metaphors using cross-linguistic awareness-raising
activities. ELT Journal 51: 352-360.
Francis, W. N. & H. Kucera
1982 Frequency analysis ofEnglish usage: Lexicon and grammar. Boston:
Houghton Miffiin Company.
Gibbs, Raymond W.
1990 Psycholinguistics ofidiomaticity. Cognitive Linguistics 1-4: 417-51.
1994 The poetics ofmind. Cambridge: Cambridge University Press.
Gibbs, Raymond W. & J. O'Brien
1990 Idioms and mental imagery: The metaphorical motivation for idiomatic
meaning. Cognition 36: 35-68.
K6vecses, Zoltan
1990 Emotion concepts. New York: Springer Verlag.
2000 Metaphor and emotion. Language, culture, and body in human feeling.
Cambridge: Cambridge University Press.
K6vecses, Zoltan & Peter Szab6
1996 Idioms: A view from cognitive semantics. Applied Linguistics 17:
326-355.
Lakoff, George
1987 Women, fire, and dangerous things. Chicago: Chicago University
Press.
Lakoff, George & Mark lohnson
1980 Metaphors we live by. Chicago: Chicago University Press.
Langacker, Ronald W.
1987 Foundations of cognitive grammar, Volume 1. Stanford: Stanford
University Press.
Lennon, Paul
1996 Getting 'easy' verbs wrong at the advanced level. IRAL XXXIVI1:
23-36.
Lewis, Michael
1993 The Lexical Approach. London: Language Teaching Publications.
Low,G.
1988 On teaching metaphor. Applied Linguistics 9: 25-49.
MacLennan, Carol H. G.
1994 Metaphors and prototypes in the learning and teaching of grammar
and vocabulary. IRAL XXXIIIl: 97-110.
Norvig, P. & George Lakoff
1987 Taking: A study in Lexical Network Theory. Berkeley Linguistics
Society 14: 195-206.
Dictionaries
1. Introduction
Language learning is one of the most complicated feats that human beings
accomplish. Any number of very real reasons exist as to why L2 learning
presents tremendous challenges. However, instructed L2 learning has been
further complicated by the fact that important elements of systematicity that
exist in language have not been appropriately captured by the pedagogical
grammars which underlie modem foreign language teaching textbooks and
materials. For instance, lexical classes, such as English prepositions, are
represented in the grammars (and the textbooks based on them) in piece-
meal fashion. When students (and their teachers) encounter varying uses of
these forms, the systematic relations between the multiple uses remain
unexplained. For example, traditional analyses have not offered any expla-
nations for why the four different meanings found in the sentences in (la-d)
are all associated with the form over:
(1) a The picture is over the mantle
b The teller at the central bank switched the account over to a local
branch
c The film is over
d Arlington is over the river from Georgetown
In sentence (la) over is interpreted as roughly 'located higher than'; in sen-
tence (lb) over is interpreted as roughly 'transferred'; in sentence (lc) over
is interpreted as roughly 'completed or finished'; and in sentence (ld) over
is interpreted as roughly 'on the other side of'. Such varying meanings are
typically presented, if they are addressed at all, as an unorganized list of
unrelated meanings that are accidentally coded by the same phonological
form. This results in a fragmented picture of the lexical class, leaving the
learner with the impression that the various uses are arbitrary. Indeed, learn-
ers of English as a second language and many teachers of ESL have noted
that acquiring the semantics of English prepositions is very difficult (e.g.,
Celce-Murcia & Larson-Freeman, 1998). In spite of the recognized difficulty,
Our first basic assumption is that the multiple meanings associated with each
preposition form a principled polysemy network organized around a central
sense, rather than a list of unrelated meanings. Two lines of argumentation
support this assumption.
First, work in experimental psychology (e.g., Rosch, 1975) has estab-
lished that humans organize their mental representations of categories around
a central exemplar that can be represented at various levels of abstraction or
generality (lohnson-Laird, 1983). Cognitive linguists (e.g., Dewell, 1994;
Langacker, 1987, 1991, 1992; Lakoff, 1987; Lakoff and Johnson, 1999;
Taylor, 1995; and Vandeloise, 1991) have extended this understanding of the
general organization of human cognition to the mental lexicon. Their work
offers strong evidence that the mental lexicon is not organized like a dic-
tionary in which each meaning associated with the same phonological form
represents an unrelated word. Rather, lexical items are better understood as
forming natural categories that participate in organized semantic networks,
or polysemy networks, organized around a central sense. Work in psycho-
linguistics (e.g., Sandra and Rice, 1995) offers empirical support for this
position.
The second line of argumentation in support of polysemy networks rep-
resents the view that the basic purpose for humans using language with each
other is communicative in nature. As a result, in naturally occurring com-
municative events, lexical items occur in sentential context, not in isolated,
citation form. Assuming that a lexical item is initially used to indicate one
established meaning, we posit that a speaker attempting to communicate
with a listener would use that lexical item to mean something new or different
from the established meaning only if they believed the listener had a reason-
able chance of understanding the new meaning. This understanding presum-
ably would come from inferences arising from the situated or contextualized
use of the lexical item as it occurs in the ongoing discourse. This suggests
that the additional meanings that have come to be associated with over orig-
inally arose from situated uses and the inferences that were derivable from
context. With repetition across a number of similar contexts, the inferences
come to be independently associated with the lexical form as additional
senses; following Traugott (1989) we term this process of extending mean-
ing pragmatic strengthening.
To summarize, our first assumption is that the multiple meanings associ-
ated with a preposition are not accidental, but rather that they are related to
each other in systematic ways represented by an organized semantic network.
ground element the Landmark (LM). We take this basic spatial relation
associated with each preposition as the central sense from which the various
additional meanings have ultimately been derived.
Conceptual content can be abstracted away from recurring spatial scenes,
giving rise to a highly abstract and schematized representation, which we
tenn a proto-scene. A proto-scene can be equated with the primary meaning
associated with a particular preposition, and thus includes infonnation relat-
ing to the TR and LM, as well as the spatial relation mediating the two. As
proto-scenes are idealized, they do not contain detailed information about
the nature of either the TR or the LM, nor detailed metric infonnation con-
cerning notions such as the exact shape of the LM or the degree of contact
between the TR and LM.3
•
Figure 1. The proto-scene for over
In our full model presented in Tyler and Evans (2001; cf. Tyler and Evans
2003), we introduce several inferencing strategies. For present purposes we
will discuss only one.
As a default, speakers assume that all elements in a conceptual spatial
scene are subject to real-world force dynamics (Tal my, 1988, 2000).
Vandeloise (1991) discusses this in terms of a naIve theory of physics that
applies to how humans conceptualize spatial relations and how they use lan-
guage to express those conceptualizations. In other words, as listeners are
interpreting utterances, they assume the objects being discussed are subject
to force dynamics such as gravity.
certain perspective ... " (1987: 39). That is, the physical vantage point on a
spatial scene will determine how we conceptualize that scene, and no two
vantage points offer the same view. Hence, as the vantage point changes,
the exact interpretation ofthe scene changes.
Consider the following example. In a scene in which a large cloth is
positioned in relation to a table so that the cloth covers the top of the table,
the scene can be construed by focusing on contact between the cloth and the
table. In this case, the scene is likely to be coded in English by the sentence:
The tablecloth is on the table. Alternatively, the relationship between the
cloth and the table can be viewed as the cloth hiding or obscuring the table
from the observer's view. In this case, the scene might be coded as: The
cloth is over the table. A less typical, but perfectly acceptable view would
be to place the table in focus, in which case the coding would be something
like: The table is under the tablecloth. Hence, the same basic scene affords
several distinct ways of being viewed and interpreted.
Although this example involves changes in lexical items in order to
signal a change in vantage point, shifts in vantage point are not necessarily
coded by changes in lexical items. We will see several examples of this in
the following section.
4.1. Overview
thus creating momentum; and 3) The preposition over prompts for a partic-
ular conceptualized spatial relation between the TR (the cat) and the LM
(the wall, which is understood as a barrier which the cat must overcome if
its forward motion is to continue),6 as depicted in the proto-scene in figure l.
When interpreted within the context of the sentence, it tells us that at some
point in its jump, the cat was located higher than the wall.
The diagram in figure 2 captures the conceptualization that might arise if
the interpretation of the utterance were based solely on the information pro-
vided by the lexical items.
•
A
. I ..
......./ .........................................................
f\~
A C
Figure 3. Schematization of normal interpretation, i.e., one involving inferences, of
sentences of the type: The cat jumped over the wall
Clearly, this is not the normal understanding of the full motion the cat would
be expected to engage in when we interpret the sentence. Rather, native
speakers normally interpret the sentence to mean something like the cat fol-
lowed a trajectory that approximates the diagram in figure 3, in which the
cat comes back down to earth, ending its jump at approximately point C.
What we want to emphasize is that nothing in the linguistic information
provided in the sentence The cat jumped over the wall specifically codes
point C. In other words, point C is inferred. The question is how this infer-
ence arises.
We argue that the interpretation of this utterance, including the inference
of point C, comes from integrating our knowledge of: 1) the real world (for
instance, our knowledge of the action of jumping which involves an ani-
mate entity creating enough momentum to push itself off the ground and
propel itself to a position higher than the wall, and our knowledge of cats -
that they cannot stay suspended indefinitely in space the way, say, a hum-
mingbird can; 2) the key spatial configuration between the TR and LM
coded by over which tells us that at some point in its movement, the cat was
positioned higher than the wall and that the wall represents a potential
obstacle to the eat's forward motion; and 3) the force dynamics of gravity
and momentum which tell us that the cat, having reached point B, must
come back to earth, point C. Thus we argue that full interpretation of the
sentence The cat jumped over the wall crucially involves the inference that
the cat ends its jump at point C.
Furthermore, we propose that repeated observations of entities engaging
in similar motion (i.e., motion that involves the entity pushing off from a
starting point, reaching a point in its movement in which it is located higher
than a LM, and then returning to ground at point C) and exposure to utter-
ances which prompt for conceptualizations of entities involved in such
motion (for instance, The girl stepped over the branch in the path, The rab-
bit hopped over the stone, The horse jumped over the fence) result in a
highly abstract schematization being established in memory. The diagram in
figure 3, which we call the A-B-C trajectory, constitutes an attempt to repre-
sent this schematization.
We hypothesize that repeated encounters with utterances involving a
particular preposition, here over, and a particular inference, that the motion
that the TR engages in involves point Q, can result in the inference itself
becoming a distinct meaning associated with the lexical item or can give
rise to secondary inferences which become distinct meanings associated
with the preposition (Traugott, 1989).
In essence, we are arguing that the inevitable inferences that occur as a part
of normal, everyday interpretations of prepositions, as they occur in senten-
tial contexts, provide a powerful mechanism for extending the meanings
associated with a preposition. To illustrate this point, we will consider three
distinct meanings associated with over, which appear to arise from natural
inferences that result from interpreting sentences which involve an A-B-C
trajectory. We saw these three distinct meanings - transfer, completion, and
on-the-other-side - illustrated at the beginning of the paper. They are
repeated again in the examples in 2-4. The diagrams in figures 4-6 are
meant to represent the schematized spatial scene that is prompted for by
each of these distinct senses associated with over. In each case, they ulti-
mately arise from the inference of C in the A-B-C trajectory, but have been
changed in particular ways in line with the two cognitive principles dis-
cussed previously. The diagrams in figures 4-6 do not look exactly like the
diagram of the A-B-C trajectory in figure 3. Each diagram in figures 4-6
reflects some change or addition to the original spatial scene which is
important in the ultimate establishment of the distinct meanings associated
with over.
A key point we will attempt to convey is that at first glance, and if we
only attend to propositional definitions, there appears to be little relation
among the three meanings: transfer, completion, and on-the-other-side.
However, if we focus on the spatial scene prompted for by the preposition
as it occurs within sentential context and on the inferences that inevitably
arise during everyday interpretation of utterances, systematic relations
among these distinct interpretations reveal themselves. With all three of
these distinct meanings, the original spatial configuration of the TR being
higher than the LM is no longer active. The crucial point we are making is
that the inference of C arose from interpreting a sentence that does involve
the proto-scene.
(2) The teller at the central bank switched the account over to a local
branch.
...............................
«(/ I
Figure 4. Transfer sense: The TR has been transferred from the left side of the imped-
iment to the right side, as represented by the dark sphere which is in focus.
that when the TR lands at point C, the process the TR is involved in is com-
pleted. In our interpretation of the film, we understand that a process is
completed, i.e., the viewing or showing of the film is completed. 7
The diagram in figure 5 represents the conceptualized spatial scene
prompted for by this utterance:
I
Figure 5. Completion sense: The dark sphere on the left represents the location of
the TR at the beginning of the process. The large sphere on the right,
which is in focus, represents the end-point or completion.
The sphere on the left represents the location of the TR at the beginning of the
action or process. The large sphere on the right is highlighted and represents
the location of the TR at the completion of the action or process. A key to
this interpretation is that the end point of the trajectory is highlighted or
given special focus. As we noted with the transfer sense, highlighting is one
of the changes in vantage point that has been identified in the CL literature,
hence establishment of the completion sense, with highlighting on point C,
involves a change in vantage point from the original scene depicted by the
A-B-C trajectory. As with all senses related to the A-B-C trajectory, point C
is taken to arise from an inference involving our knowledge of force
dynamics. Thus, we again see that the principle that a scene can be viewed
from many vantages, in conjunction with background knowledge of force
dynamics, combine to give rise to a new interpretation.
It is important to point out that in this use of over the focus is on point C,
the point at which the action is completed. The spatial configuration associ-
ated with the central sense of over is no longer strictly associated with this
sense. The location, point C, has been re-interpreted as providing information
about the action or process, not the spatial relation between the TR and the
ting a process and as such is now acting as an adverb. Through repetition
and entrenchment in memory - the process we are identifying as pragmatic
strengthening - the repeated inference of "completion" has come to be inde-
pendently associated with over as a distinct sense.
« e» I•
Figure 6. On-the-other-side-of sense: The eye icon on the left represents the van-
tage point, the verticalline the impediment and the dark sphere the TR.
The vertical line represents the LM. The shaded (and hence highlighted)
sphere represents the TR, which is construed as being at the completion
point of the action. The eye icon on the left represents the vantage point,
which locates the conceptualizer as being onstage and represents a shift
from the basic A-B-C schematization.
above-and-beyond
(excess I) completion
2.B 2.C
transfer focus-of-attention
covering 4.A
3
6 reflexive
5.A
more 5.B 5.C 6.A
control preference repetition
5.A.l
over-and-above
(excess JI)
We believe that the approach to prepositions we have outlined here has the
potential to provide a number of benefits for the second language learner.
First, the model represents the various, extended senses associated with over
as being clearly motivated by a relatively small number of principles. Although
we acknowledge that some uses of over (and other English prepositions) are
bound to be idiosyncratic, especially, for instance, when they combine with
verbs in verb-particle constructions, the amount of arbitrariness is significantly
reduced under the current approach. Thus the model provides a more sys-
tematic, explanatory account of the semantics of English prepositions than
traditional approaches, cutting down considerably on the amount of arbitrari-
ness in the representation and hence reducing the need for rote learning on the
part of the second language learner. Second, because the model draws heavily
on the notion of the experiential basis of meaning and represents the extended
senses as arising from observations of the external, spatio-physical world, it
reflects the learners' own experiences with the world. Understanding the mo-
tivation behind the extended senses as experientially motivated and coherent
with the learners' own observations of the world would seem to make these
senses easier to acquire. Third, the various senses are represented as gestalt-
like conceptualizations of situations or scenes which are systematically con-
nected, rather than a series of discrete dictionary-type definitions strung to-
gether in a list. The systematic connections for over are modeled in the graphic
representation of the network, as illustrated in figure 7. Such graphic repre-
sentations of polysemy networks provide visual rubrics that may be useful
presentational tools for the language teacher and useful aids for the second
language learner. Finally, the constrained, principled nature of the model
would seem to provide a solid foundation for the learners from which to infer
the meanings of unfamiliar uses of over when they are encountered in context.
In the remainder of this section we offer a few suggestions concerning
how the proto-scene and two of the extended senses might be taught. These
ideas and materials have been piloted in a small, quasi-experimental class-
room intervention (Winke & Kim, 2002). These lessons are aimed at inter-
mediate-level English language learners who presumably have already been
exposed to some version of the central sense of over. The teaching activities
themselves and their sequencing are motivated by the model we have out-
lined. They draw on the notions that observations of the external, spatio-
physical world provide cognitive framing for the internal, conceptual world
and that cognitive representations of observations of events in the word
involve a scene complete with participants engaged in the event.
c. Give several examples, again using pictures, flip books, etc. E.g., 'The
jump is over'.
d. Then explain that once over became associated with completion of
physical movement, it could be extended to mean completion in gen-
eral. Give several examples of non-physical uses such as 'Class is
over'.
4. Move to the Transfer sense.
a. Ask a student to come to the front of the room. Stand on one side of
the desk (point A); ask the student to stand on the opposite side of the
desk (point C). Throw some large, silly object (like a nerf ball) over
the desk to the student. Repeat the throw, but before you throw the
object ask, 'Who has the X?' Toss the object, when the student has
caught the object ask, 'Now who has the X? How did the X get from
me to student? I threw it OVER the desk. I tossed it OVER the desk.'
b. Emphasize that: 1) because when an object moves from point A to
point B to point C, the object is transferred from A to C; and, 2)
because English speakers use over to describe the whole A-B-C scene,
over has taken on the additional meaning of 'transfer'. Reinforce with
several examples of physical transfer over an obstacle while intro-
ducing common phrases such as hand over, pass over, toss over, etc.
c. Explain that once over was commonly used to describe transfer of
physical objects, it was extended to indicate transfer in general: sign
over, turn over, win over, take over, etc. E.g.,
- The Beatles immediately won the hearts of millions of teenagers.
- The Beatles eventually won over the hearts ~f many of their parents
as well. {Note how the use of over raises the implication of an obsta-
cle that had to be overcome.}
- After long debate, George Bush managed to win over a few govern-
ments to his position on Iraq.
7. Conclusion
We have analyzed the mUltiple senses associated with each English preposi-
tion as forming a polysemy network organized around a central sense, the
proto-scene, which is made up of a TR and a LM in a specified spatial con-
figuration and a functional element. Each proto-scene is understood to con-
stitute the primary meaning representation associated with a particular
preposition, from which additional meanings have been systematically
derived. Thus, each preposition and the multiple uses associated with it are
Notes
1. In the full analysis of over, 14 senses are identified, (cf. Tyler and Evans, 2001).
Moreover, we emphasize that the model is based on a thorough analysis of 17 of
the most commonly occurring English prepositions (cf. Tyler & Evans, 2003).
2. Langacker (1987) argues "sensory imagery is a real phenomenon whose role in
conceptual structure is substantial. We can plausibly suppose that a visual image
(or a family of such images presupposing different orientations and levels of
specificity) figures in our knowledge of the shape of an object; and certainly one
aspect of our conception of a trumpet assumes the form of an auditory image
representing the sound it makes" (pp. 111). He emphasizes that a commitment to
the importance of sensory imagery in the shaping of conceptual structure does
not imply a position that sensory imagery is an exclusive or essential facet of all
meaning of linguistic expressions. Neither should sensory images be confused
with the naIve view that a sensory or even a conceptual visual image is analogous
to a photograph or a picture. As the experimental psychologist Kosslyn (1980)
argues, "Image representations are like those that underlie the actual experience,
but in the case of mental imagery these representations are retrieved or formed
from memory, not from immediate sensory stimulation" (p. 18).
3. In our full model, the notion of a functional element plays a crucial role. We
hypothesize that in addition to the spatial configuration between a TR and a LM,
the concept prompted for by a preposition also involves a functional element
(Tyler and Evans, 2003; Vandeloise, 1991). In the case of in, for example, the
functional element involves the notion of containment. Johnson (1987), for in-
stance, has argued that the functional element of containment includes location,
confinement, protection, and potential obscuring of the element(s) being con-
tained. For instance, if a young child is in a playroom, the caretaker knows where
the child is located, the actions of the child are limited to those which can take
place within the space of the playroom, the child is protected from certain out-
side threats (e.g., the hot stove in the kitchen), and, for the most part, the child is
physically obscured from entities outside the playroom. The container and its
interior region also form the physical environment, which surrounds the TR. In
the case of the child in the playroom, the interior region of the room largely
determines the temperature, lighting, ambient sounds, etc, in other words, the
general physical environment which surrounds the child.
Our analysis has revealed that the functional element is key to appropriately
characterizing the distinction between the prepositions over and above. However,
the points made in the present discussion do not crucially refer to the functional
element.
4. The TR being potentially within reach of the LM allows over to depict spatio-
physical configurations in which there is contact between the TR and the LM as
well as those in which there is no contact. This is a crucial distinction between
over and above (whose functional element is distal in nature). The difference in
References
Bloomfield, Leondard
1933 Language. New York: Holt, Rinehart and Winston.
Brugman, Claudia
1981 The story of over. MA Thesis. Linguistics Dept., UC Berkeley.
Published [1988] as The story of over: Polysemy. semantics and the
strncture of the lexicon. New York: Garland Press.
Celce-Murcia, Marianne and Diane Larsen-Freeman
1998 The Grammar Book: An ESLlEFL Teacher s Course. Rowley, Mass.:
Newbury House.
Chomsky, Noam
1995 Categories and transformations. In The minimalist program, 219-394.
Cambridge, MA : MIT Press.
Dewell, Robert
1994 Over again: Image-schema transformations in semantic analysis.
Cognitive Linguistics, 5, 351-380.
Frank, Marcella
1972 Modern English: A Practical Reference Guide. Englewood Cliffs, NJ:
Prentice-Hall.
Grady, Joseph
1997 Foundations ofmeaning: Primary metaphors and primary scenes. Lin-
guistics Dept., University of California at Berkeley. Ph. D Dissertation.
1999 A typology of motivation for conceptual metaphor: Correlation vs.
resemblance. In G. Steen and R.Gibbs (eds.). Metaphor in Cognitive
Linguistics (pp. 27-47) Philadelphia, PA.: John Benjamins.
Green, Georgia
1989 Pragmatics and Natural Language Understanding. Hillsdale, NJ:
Lawrence Erlbaum Associates.
Grice, H. P.
1975 Logic and conversation. In P. Cole & J. Morgan (eds.) Syntax and
Semantics. vol. 3, 41-58. New York: Academic Press.
Johnson, Mark
1987 The Body in the Mind: The Bodily Basis of Reason. Thought and
Imagination. Chicago: Chicago University Press.
Johnson-Laird, Philip
1983 Mental Models. Cambridge, MA: Harvard University Press.
Kosslyn, Stephen
1980 Image and Mind. Cambridge, MA: Harvard University Press.
Kreitzer, Anatol
1997 Multiple Levels of Schematization: A Study in the Conceptualization
of Space. Cognitive Linguistics, 8,4,291-325.
Lakoff, George
1987 Women. fire and dangerous things: What categories reveal about the
mind. Chicago, IL: Chicago University Press.
Langacker, Ronald
1987 Foundations of Cognitive Grammar, vol. 1. Stanford, CA.: Stanford
University Press.
1992 Prepositions as grammatical(izing) elements. Leuvenese Bijdragen,
81,287-309.
Lindner, Susan
1981 A lexico-semantic ana~vsis ofEnglish verb particle constructions with
out and!l/2.. Doctoral dissertation. Linguistics Dept., UC San Diego.
Distributed [1983] by the Indiana University Linguistics Club.
Lindstronberg, Seth
1998 English prepositions explained. Amsterdam: John Benjamins.
Mandler, Jean
1988 How to Build a Baby: On the Development of an Accessible Repre-
sentational System. Cognitive Development, 3, 113-136.
1992 How to Build a Baby: H. Conceptual Primitives. Psychological
Review, 99, 4, 587-604.
1996 Preverbal Representations and Language. In P. Bloom et al. (eds.)
Language and Space, 365-384. Cambridge, Mass.: The MIT Press.
Rosch, Eleanor
1975 Cognitive representations of semantic categories. Journal of Experi-
mental Psychology: General, 104, 192-233.
Sandra, Dominiek & Sally Rice
1995 Network analyses of prepositional meaning: Mirroring whose mind-
The linguist's or The language user's? Cognitive Linguistics, 6, 1,
89-130.
Sperber, Daniel and Dierdre Wilson
1986 Relevance: Communication and Cognition. Cambridge, MA: Harvard
University Press.
Talmy, Leonard
1978 Figure and ground in complex sentences. In J. Greenberg (ed.).
Universals in human language, vol. 4, 625-649. Stanford, CA.:
Stanford University Press.
1988 Force dynamics in language and cognition. Cognitive Science, 12,
49-100.
2000 Toward a Cognitive Semantics. Cambridge, MA: MIT Press.
Traugott, Elizabeth
1989 On the rise of episternic meanings in English: An example of subjec-
tification in semantic change. Language, 65,1,31-55.
Tyler, Andrea and Vyvyan Evans
2003 The Semantics of English Prepositions: Spatial, Embodied Meaning
and Cognition. Cambridge: Cambridge University Press.
Vandeloise, Claude.
1991 Spatial prepositions: A case study in French. Chicago: The University
of Chicago Press.
Winke, Paula and YiYoung Kim
2002 Poster presented at Second Language Research Forum. Toronto,
Canada. (October 2002).
radial category 8
retention 8,212-216,220-222,224- Usage event 55,179-180
226,249 usage-based 4-6,51, 54-56, 77, 122,
124,258
second language acquisition 6, 13, 18, usage-based approach 5, 51, 54-56,
26, 115, 119, 121-122, 130, 135, 77,258
137-139,170,172
semantic extension 4, 8
semantic prime 144 vocabulary 8, 22, 26, 100, 112-113,
Semantics of English prepositions 145,149,157,159,181-182,211-
257-258,273 215,220-222,224-226,228,233-
similarity 6, 77-78, 81-82, 109, 158, 234,236,248-250
178
since 195-196, 200, 204-208 ways of speaking 144-145,150
spatial scenes 261-263 when 195-203, 205,208
speaker construa151, 55-56 while 195-196, 201, 204-206, 208