Content uploaded by Jeannette Littlemore
Author content
All content in this area was uploaded by Jeannette Littlemore on Mar 16, 2018
Content may be subject to copyright.
Uncorrected proofs - John Benjamins Publishing Company
Metaphoric competence
intherstandsecondlanguage
Similarities and dierences
Jeannette Littlemore
University of Birmingham
. Introduction
In recent years there has been a growing interest in the teaching of metaphor to second
language learners (Danesi, 2008; Holme, 2004), which has partly been due to an apprecia-
tion of the fact that the ability to understand and use metaphors in the target language is
likely to make a substantial contribution to second language prociency. An ability to un-
derstand and produce metaphor can help learners to develop their sociolinguistic, illocu-
tionary, grammatical, discourse and strategic competence (Littlemore and Low, 2006a,b).
Despite this growing interest, there has been surprisingly little research into the extent to
which language learners are able to transfer their metaphor interpretation and produc-
tion skills (or behaviour patterns) from their mother tongue (L1) to the target language
(L2). e only study of which I am aware is Johnson (1989) who looked at metaphor
interpretation skills in a group of bilingual children. e rst language of all the children
was Spanish but they all attended English-medium schools. She found that complexity of
the interpretations oered in English correlated signicantly with the complexity of inter-
pretations oered in Spanish. ese ndings suggest that metaphor interpretation skills
constitute an individual dierence variable. Despite these intriguing ndings, no further
studies have looked at how metaphoric competence develops in the bilingual lexicon, or at
how it contributes to the sort of linguistic multicompetence that L2 users acquire (Cook,
2002). Information on this matter would help language teachers to decide upon the extent
to which it is necessary to teach metaphor skills in the target language, and the extent to
which they can safely assume that they consist of cognitive skills that learners will simply
transfer from their L1.
In order to investigate this issue, an exploratory study was conducted to investigate
whether or not there is a relationship between metaphor interpretation and production
behaviour in the L1 and the L2. e focus of this study was on decontextualised linguis-
tic metaphors, and its ndings do not tell us anything about a language learner’s ability
to understand and produce either conceptual metaphors (Lako and Johnson, 1980), or
Uncorrected proofs - John Benjamins Publishing Company
Jeannette Littlemore
contextualised linguistic metaphors (Cameron, 2003), but they do tell us something about
the relationship and dierences between L1 and L2 metaphoric processing and/or ability.
. e development of metaphor interpretation
and production skills in the rst language
e ability to understand and produce metaphors develops at a relatively early stage in
rst language acquisition. Research on children has suggested that both the production
and comprehension of linguistic metaphors are skills that increase with age. Gardner et
al. (1974) carried out an experiment into children’s capacities to both create and appreci-
ate metaphors. In the rst part of their experiment, the children were instructed to think
of appropriate metaphorical endings for a list of twelve very short stories. In the second
part the children were told to choose the “best” metaphor from a list of four metaphors
containing a novel metaphor, a literal comparison, a conventional metaphor and an inap-
propriate metaphor. In each case, they were told that marks would be gained for novelty.
Gardner et al. observed a tendency, increasing with age, towards novel metaphoric
production and preference. ey also found that the highest percentages of novel met-
aphors were produced by the youngest participants and the oldest participants. Where
these two groups diered was in the tendency of young participants to produce metaphors
which were highly original but inappropriate or nonsensical. eir ndings led them to
claim of the oldest participants that (ibid.: 138–139):
e greater cognitive sophistication of the older participants enables them to appreciate
the links between the disparate realms... e oldest participants generally display an ex-
plicit appreciation of metaphor, oen prefer metaphoric endings, have some capacity to
produce vivid comparisons if not always to shi across domains, and the critical ability
lacking among the youngest participants. ey alone possess the metalinguistic air of re-
ecting upon the words of others, and of considering what is appropriate in a given context.
However, the idea that the ability to understand metaphor does not develop until late
childhood has been contested. Waggoner and Palermo (1989) asked thirty two children
of three age levels (ve, seven and nine years) and thirty two college students to interpret
and explain metaphors describing love, hate, happiness, sadness, anger, and fear. eir re-
sults showed that even the youngest children demonstrated some ability to interpret meta-
phors, although they could not explain their interpretations. e appreciation of meta-
phor by young children is also observed by Elbers (1988) who notes children’s tendencies
to comment on similarities from the age of two, although they cannot explain the meaning
of these similarities. e general consensus would seem to be that children do possess
some ability to perceive metaphor from a very early age, but that this ability increases as
they get older and that older children are more able to explain and produce metaphor1.
Researchers have wondered whether this change is due to a developmental increase
in domain-specic knowledge, linguistic ability or general cognitive capacity. Billow
(1975) carried out an experiment, the results of which led him to the tentative suggestion
that metaphoric competence was related to cognitive development. In order to test this
Uncorrected proofs - John Benjamins Publishing Company
Chapter 15. Metaphoric competence in language
further, Johnson (1991) examined the metaphor interpretation skills of children with
diering levels of language prociency (as measured by the oral language subscale of the
Woodcock Language Prociency Battery, – Woodcock, 1980) and mental capacity (as
measured by the Figural Intersections Test – Pascual Leone and Burtis, 1975). She found
that mental capacity was a more important indicator of metaphor interpretation ability
than language prociency.
Support for the idea of metaphoric development as a cognitive phenomenon has
also been demonstrated by Kogan (1983) who administered the Metaphoric Triads Task
(MTT) to children of dierent ages. e MTT is a non-linguistic test comprising twenty
nine triads of pictures. Each triad oers three pairing possibilities. For example, one triad
oers pictures of a toddler on the grass, a brightly coloured watering can, and a rosebud.
e following pairing links are possible:
Toddler and watering can: e toddler can play with the watering can;
Watering can and rosebud: One can sprinkle water on the rosebud with the watering can;
Toddler and rosebud: Both are at an early stage of development
e third link is considered to be the metaphorical link. e children are asked to make as
many pairings as possible between the pictures. As this test has no overt linguistic element
it is hypothesized to be a test of pure cognitive development. Using this test, Kogan found
a progressive increase in metaphoric comprehension across an age span extending from
kindergarten to young adulthood.
e idea that the ability to interpret metaphors involve more generic cognitive op-
erations provoked two attempts to investigate various Piagetian operations as possible
cognitive prerequisites for metaphoric thinking. Billow (1975) investigated the relation-
ship between combinational reasoning (one of Piaget’s formal operations) and metaphoric
comprehension, though he was unable to nd clear evidence for a relationship. Cometa
and Eson (1978) found that intersectional classication (one of Piaget’s concrete-opera-
tional operations) did correlate with an ability for metaphoric comprehension. us it
would seem that the ability to understand metaphors is a cognitive phenomenon which
develops during the concrete operations period (four-to-eleven year olds). is nding is
in accordance with all the above experiments which suggest that metaphoric competence
is a cognitive skill which develops during childhood. Clearly then, most types of meta-
phoric competence involve cognitive skills, and they develop with age.
More recent work on metaphor comprehension in the rst language has also sug-
gested that metaphor interpretation and production reects underlying general cognitive
processes (see Evans and Green, 2006). ese processes include the activation of relevant
domain knowledge (Giora, 2003), imagery (Li, 2002), episodic memory (Bottini, et al.,
1994), analogical reasoning (Paivio and Walsh, 1993), categorization (Glucksberg et al.,
2001), the use of context (Gibbs, 1994), associative uency (Johnson and Rosano, 1993),
and conceptual blending (Fauconnier and Turner, 1998). Although there is very little
universal agreement over which of these processes are involved in metaphor production
and comprehension, or how they interact, the fact that they are all cognitive processes
suggests that people may exhibit individual dierences in these areas that are relatively
Uncorrected proofs - John Benjamins Publishing Company
Jeannette Littlemore
stable across languages. us a tendency/ability to understand and produce metaphor in
the L1 may be related to the same tendency/ability in the L2. at is to say, people who
are good at understanding and producing metaphor, or who like to do so, in the L1, may
also display the same tendency in the L2, as far as their linguistic resources permit. In the
following section I review the limited amount of research that has investigated L1 and L2
metaphor comprehension, interpretation and production.
. e relationship between metaphor interpretation
and production skills in the L1 and the L2
ere have been a few studies comparing the metaphoric competence of L1 speakers
with that of L2 speakers. Both the Trosborg (1985) metaphoric production study and the
Johnson and Rosano (1993) metaphoric interpretation study which are described above
were carried out with native speakers as well as to language learners. e only overall
dierence observed by Trosborg between L2 students and native speakers was that the
native speakers had a higher tendency than all the other groups to produce conventional
metaphors. is could be because the native speakers were simply more aware of the con-
ventional metaphors than the other groups. e dierence would therefore not seem to
be due to variations in cognitive processing, but simply to gaps in the knowledge of the L2
learners. Johnson and Rosano found no group dierences for either of the tests of meta-
phor interpretation. Finally Johnson (1996) found that L2 prociency and L2 metaphor
interpretation abilities were unrelated, and that a better predictor of L2 metaphor inter-
pretation abilities was the participant’s L1 metaphor interpretation abilities. It is there-
fore reasonable to hypothesise that L1 and L2 metaphoric competence are very probably
related. However, neither of these studies looked at performance in the tests by the same
population using their L1 and their L2. Moreover, they did not investigate dierent types
of metaphoric competence. Given the range of functions performed by metaphor, it is
important that language learners can understand it. ey need to be able to understand it
relatively rapidly as an inability to do so may hold up the conversation and may thus be a
source of frustration on both sides. It is also important that they can interpret it at dierent
levels as one of the main advantages of using metaphor is that it can say several dierent
things at once. Of less immediate necessity, but still relatively important is the ability to
produce metaphor in the target language.
e aim of the study described in this article was to explore metaphor comprehen-
sion and production skills in the L1 and the L2 in order to determine whether there is a
relationship between them, and whether participants are generally better at dealing with
metaphor in their L1 or their L2. Four dimensions of metaphoric competence were ex-
plored. ese were: tendency to nd meaning in metaphor, speed in nding meaning
in metaphor, ability to identify multiple interpretations for a given metaphor, and novel
metaphor production.
Uncorrected proofs - John Benjamins Publishing Company
Chapter 15. Metaphoric competence in language
. e study
e study was designed to investigate two research questions:
1. If a student displays a high level of metaphoric competence in the L1, will they also
display a high level of metaphoric competence in the L2?
2. Are levels of L1 metaphoric competence higher or lower than levels of L2 metaphoric
competence?
In order to answer these questions, four tests were devised. ese were: a test of one’s ten-
dency to nd meaning in metaphor, a test of speed in nding meaning in metaphor, a test
of one’s ability to identify multiple interpretations for a given metaphor, and a test of novel
metaphor production. e participants were 82 upper-intermediate French-speaking uni-
versity students of English studying at a university in Belgium. All of the participants were
in their second year of a degree in Languages and Linguistics. Seventy-ve of the partici-
pants were female and seven were male. Before participating in the tests the participants
were informed by their class teachers that they would be asked to participate in a piece of
research into language learning strategies. ey were told that participation was optional
but that the tests would take place during class time. Out of a total possible of eighty ve
students, eighty two agreed to complete the tests and signed consent forms.
e tests were administered in two sessions, a two-hour session was based in a lan-
guage classroom and language laboratory, and one one-and-a-half-hour session was based
in a computer suite. In each session there were approximately ten students. All students
completed the classroom-based session rst and the computer-based session second, with
a two-week period separating the two sessions. e participants asked to perform the tests
in English and in French. e format of the test was identical in both languages, although
the items varied. e tests were as follows:
. Testing the participants’ tendency to nd meaning in metaphor
and speed in nding meaning in metaphor
In order to measure one’s tendency to nd meaning in metaphor, a test was created which
was partly inspired by the test devised by Pollio and Smith (1979). In Pollio and Smith’s
test participants were given a list of sentences of the form “e _____ is a _____” and they
were asked to classify each sentence as one of the following ve types:
For example:
A typical type 1 (synthetic) sentence was “e dog is a poodle.”
A typical type 2 (analytic) sentence was “e tulip is a ower.”
A typical type 3 (contradictory) sentence was “e dog is a cat.”
A typical type 4 (anomalous) sentence) was “e mountain is a frog.”
A typical type 5 (metaphoric) sentence was “e woman is a rose.”
As the focus of interest in this study was on students’ tendencies to dierentiate between
metaphor and anomaly, it was decided that only the fourth and h categories in Pollio
Uncorrected proofs - John Benjamins Publishing Company
Jeannette Littlemore
and Smiths’ experiment would be used (metaphor/anomaly). Pre-piloting showed that
participants found it easier to make their choice from a continuum of responses rather
than to have to make a yes/no decision.
In conversational discourse containing metaphors it is important that the metaphors
are understood quickly if the communication is to remain natural. is is particularly so
in the case of foreign language discourse. If a learner is able to produce and to comprehend
metaphors at a natural speed then this is likely to make a considerable contribution to his/
her communicative competence. It was also suggested above that the speed with which
people nd meaning in metaphors reects their loose analogical reasoning skills. e rea-
soning behind this assumption was that the fast interpretation of a metaphor depends on
the speed with which a participant is able to draw analogies between its topic and vehicle.
A decision was therefore taken to measure not only the students’ tendency to nd
meaning in metaphors, but also the time required for them to make a positive response
whenever they did so. It was decided that a computer-based version of the test would per-
mit accurate measurement of the students’ response times. is is not the rst time that
a computer-based test has been used to measure the amount of time required to process
metaphors. Gregory and Mergler (1990) investigated the amount of processing time re-
quired to decide whether or not metaphors were meaningful, and the processing time re-
quired to decide whether or not metaphorical sentences made sense. ey found that par-
ticipants took signicantly longer to decide whether or not metaphors were meaningful
than they did to decide whether or not they made sense, thus highlighting the importance
of the instructional set. In the light of Gregory and Merglers’ ndings, extra attention was
paid to the clarity of the instructional set in this test.
Piloting the tests
In four pilot sessions a total number of thirty eight participants matched to the target
population completed the following computer-based metaphor test which was created us-
ing Hypercard:2
First the participants were shown a rubric (see Appendix 1) explaining what is meant
by the terms “metaphor” and “anomaly”. is rubric also told them that they were going
to be shown a series of metaphors and that their task was going to be to rate, on a scale
of one to ve, the extent to which they thought each metaphor made sense. In this rubric
it was made very clear that their task was not to say whether or not they agreed with the
metaphor. When they had read the rubric, they were told to click on a button in the bot-
tom right hand corner of the screen and at this point the test began.
roughout the test a French version of the following scale was displayed on the right
hand side of the screen:
(5) “It’s obviously a metaphor. e relationship between the two elements is clear”
(4) “e metaphor is less convincing. One can see that there is a relation but it’s not
immediately obvious”
(3) “is is the middle of the scale. You’re really not sure if it’s a metaphor or not”
(2) “ere could be a metaphorical meaning but you can’t see it”
(1) “It’s obviously an anomaly. It is not possible to nd a relationship between the two elements”
? ere are words I do not understand in this sentence
Uncorrected proofs - John Benjamins Publishing Company
Chapter 15. Metaphoric competence in language
irty three English metaphors and thirty four French metaphors were then dis-
played, one aer the other at the top of the screen, and the participants were asked to make
their decisions and to click on the appropriate part of the scale for each response. Half of
the participants were shown the French metaphors rst, and half of the participants were
shown the English metaphors rst. e French metaphors were translations of metaphors
taken from the Katz et al study. One possible weakness of the study is that the Katz et al
ratings may not have held for the French translations.
e metaphors had been pre-selected from an article by Katz et al. (1988) in which
four hundred and sixty four metaphors were normatively rated on ten scales by six hun-
dred and thirty four raters. According to Katz et al., the aim of this article was to help
researchers interested in systematic investigation of metaphoric processes control for dif-
ferent metaphor-related variables. e ten scales along which the metaphors were mea-
sured were: comprehensibility; ease of interpretation; degree of metaphoricity; metaphor
goodness; metaphor imagery; participant imagery; predicate imagery; felt familiarity; se-
mantic relatedness; number of alternative interpretation. For this test, metaphors which
scored highly on comprehensibility were chosen. ere were two reasons for the decision.
Firstly, the highly comprehensible metaphors contained vocabulary that the students were
more likely to understand. Secondly, the chosen metaphors were relatively concrete, and
concrete metaphors might be expected to provoke more imagery than abstract metaphors
in the interpreting process. Six “nonsense” metaphors were also included in this piloting
session to control for random response.
Statistical analysis revealed no random response. Furthermore the reliability of the
test was found to be high (0.88 for the English version and 0.84 for the French version).
It was therefore decided that in the nal version of the test twenty ve English metaphors
and twenty ve French metaphors would suce.
e Katz et al. scales were useful for deciding whether or not the participants actu-
ally carried out the procedure suggested in the rubric. If they had understood the rubric
as it was meant to be understood their replies would correlate most closely either with
Katz et als.’ “comprehensibility” or “semantic relatedness”. However, correlational analysis
revealed that the participants’ results correlated most closely with Katz et als.’ “metaphor
goodness” (correlation coecient = 0.76; p = 0.00). e correlations with comprehen-
sibility and semantic relatedness were only (0.62) and (0.65) respectively which means
that the participants had judged the metaphors not for comprehensibility or for semantic
relatedness, but for whether or not they thought that they were good metaphors. However
there was a high correlation both between Katz et al’s measures of metaphor goodness
and comprehensibility (correlation coecient = 0.84; p = 0.00) and semantic relatedness
(correlation coecient = 0.90; p = 0.00) indicating that the three are very closely related.
e nal version of the test
e nal version of the test was identical to the pilot version except that the students
were asked to assess the meaningfulness of only y metaphors (twenty-ve in English
and twenty-ve in French) (see Appendix 2), on a scale from one to ve. Also, only two
nonsense metaphors were le in as control items. e scale of responses was the same as
above, except that this time the following message ashed on the screen:
Uncorrected proofs - John Benjamins Publishing Company
Jeannette Littlemore
ATTENTION!
Il ne s’agit pas de juger de la qualité du lien entre les deux éléments mais de se limiter à
décider s’il y en a un ou pas!
(Remember that your task is simply to decide whether there is a relationship between the
two elements of the sentence, not to judge the quality of that relationship!)
is was to discourage the students from making judgements of “metaphoric goodness”
and to encourage them to focus on comprehensibility and semantic relatedness. Both the
types of response and the reaction times were measured for each student for each meta-
phor (= test of tendency to accept the metaphors as meaningful, and speed of interpreta-
tion). Two scores were calculated for each student: a score indicating the average response
given (on the scale from 1 to 5). is was designed to show the general tendency of the
student to nd meaning in metaphor; and a score indicating the average amount of time
taken to decide on a positive response (4 or 5).is was designed to give an indication of
the speed with which a student was able to nd meaning in metaphor.
Correlational analysis revealed that students were still making judgements based pri-
marily on “metaphor goodness” on the French part of the test (correlation coecient =
0.9; p < 0.01), but that their judgements in the English part of the test were more related to
comprehensibility (correlation coecient = 0.83; p < 0.01). is perhaps to be expected as
one rst focus, when dealing with another language is whether or not one can understand
it; judgements of quality are something of a second-order decision in that they presuppose
comprehension. e fact that the students seemed to be doing slightly dierent things in
the dierent language versions of the test is a potential weakness of the study. e implica-
tions of this are discussed in Section 6 below. e control items showed that, again, there
was no random response factor. Furthermore, the reliability of the nal version of the test
was high (0.85 for the English version and 0.83 for the French version).
. A written “metaphoric uency” test
Metaphoric uency is dened as the mean number of interpretations that a participant
can nd for a number of metaphors. It was hypothesized above that metaphoric uency
would serve as an appropriate metaphor test to involve the psychological process of as-
sociative uency. e reasoning behind this assumption was that the more associations a
participant is able to nd between the two components of a metaphor, the more interpre-
tations he/she will be able to make of that metaphor.
In the pilot test, when the students had completed the computer-based test described
above they were presented with the ve metaphors that they had accepted most readily in
each language. ey were asked to write down as many interpretations as possible for each
metaphor. e reason why they were given the ten metaphors that they had most readily
accepted was because it was thought that this way no student would be asked to interpret a
metaphor which they had previously described as “uninterpretable”. eir interpretations
were then counted and the student was duly attributed a score for “metaphoric uency”.
is was simply a measure of the average number of interpretations oered by a partici-
pant per metaphor.
Uncorrected proofs - John Benjamins Publishing Company
Chapter 15. Metaphoric competence in language
It was found, however, that giving each student ten dierent metaphors to interpret
on the basis of their results in the previous test yielded data that were dicult to com-
pare statistically. It was therefore decided that each student should receive the same ten
metaphors to interpret (see Appendix 3). ese items were selected from the piloting of
the computerised metaphoric preference test described above. e metaphors chosen for
inclusion in this part of the test were the ve English metaphors and the ve French meta-
phors that had been most readily accepted by the highest number of students. When the
students had completed the computer-based test, they were then asked to write down as
many interpretations as they could think of for these ten metaphors.
e students’ metaphor interpretations were then counted by two independent judges
in order to assess the students’ metaphoric uency. A problem arose with one of the Eng-
lish metaphors:
“A dog is a walking stick.”
Despite the fact that in the pilot session this was one of the ve most readily accepted
English metaphors and participants had no problem understanding it, analysis of data
from the nal test session revealed the presence of a vocabulary problem. A majority of
the students had not heard of a “walking stick” and thus understood this sentence to mean
that a dog is a stick that walks. Naturally this led to problems for the interpretation of the
metaphor and the item had to be eliminated. ere were therefore only four useable items
in the English version of the metaphoric uency test. is explains why, even in the nal
version of the test, although the reliability of the French version of this test was reasonably
high: 0.65, the reliability of the English version was only 0.31.
. A “metaphor production” test
A metaphoric production test was included as it was felt that it would give more free
reign to the students’ creativity than the other tests. e “metaphoric production” test
was adapted from a technique originally used by Gardner et al. (1974) to test children’s
capacity to create and to appreciate novel metaphors, and which was subsequently used
by Trosborg (1985) to investigate the metaphoric skills of foreign language students. Tros-
borg’s experiment was divided into two parts. In the rst part of the experiment, partici-
pants were asked to think up appropriate metaphorical endings for a list of twelve very
short scenarios. ey were told that credit would be given for novelty. In the second part
of the experiment, they were told to choose what they considered to be the “best” ending
from a list of four endings containing a novel metaphor, a literal comparison, a conven-
tional metaphor and an inappropriate metaphor. It was assumed that those students who
had a strong appreciation of metaphor would elect a novel response. e students were
therefore given a version of this test.
In a piloting session, it was found that the reliability of the second part of Trosborg’s
test was very low (Cronbach’s alpha = 0.32), whereas the rst part yielded relatively high
reliabilities. It was therefore decided to administer only the rst part of the original test.
Furthermore, it was decided that reliable results could be obtained by using only eight
Uncorrected proofs - John Benjamins Publishing Company
Jeannette Littlemore
items (instead of the original twelve). In order to select the eight sentences to be com-
pleted, twelve sentences were piloted on a matched group of language students. Only those
items that caused no confusion, accounted for statistical variation and were statistically
reliable were accepted for inclusion in the nal eight items. e students were therefore
given eight uncompleted sentences in English and eight in French (half the students were
given the English sentences rst and half the students were given the French sentences
rst) and were told to complete them imaginatively (see Appendix 4).
e test was adapted slightly for another reason. In the Trosborg study the sentences
to be completed all took the form of similes (“for example “it was as quiet as...”). Re-
searchers have suggested that similes probably require slightly dierent thought processes
frommetaphors. Ortony (1979 and 1980) suggests two ways in which similes might be
seen to dier from metaphors. In his 1979 article he claims that the “nonliteral similarity”
contained in metaphors is dierent from the “literal similarity” contained in similes. In his
1980 article he goes on to propose that metaphors violate what he refers to as “the sincer-
ity postulate” (ibid.: 76). is is a concept derived from one of Grice’s (1975) maxims and
which basically means “try to mean (literally) what you say and imply”. Similes do not vio-
late this postulate and therefore might, according to Ortony, be regarded as literal rather
than metaphorical language. erefore there is a case for not treating them in the same
way as we treat metaphors. ese dierences led Kogan (1983: 660) to claim that “there is
clearly no basis for asserting that the ability to reason by analogy is a precursor of meta-
phoric thinking”. As the aim of this study was to examine metaphoric processes, the test
was adapted to provoke metaphors rather than similes. Aer having seen two examples,
one in English and one in French, the students were given sixteen very short expressions
(eight in French and eight in English) and were asked to produce endings they liked and
found right for these expressions. ey were asked to complete the sentences as creatively
as possible. For example the students were asked to complete the sentence:
e lake was a shining ................................... at the bottom of the valley.
Responses were scored as follows:
1 = inappropriate ending
2 = literal ending
3 = conventional metaphor ending
4 = novel metaphor ending
Endings were considered to be “novel” only if they met at least one of the following criteria
(adapted from Gardner et al., 1974):
a. e topic was projected onto a sensory domain where it was not literally applicable,
and the resulting metaphor was not a familiar English or French saying. (For example:
Dr. Livingstone had been walking across the Sahara for ve days without any water.
His throat was beginning to feel as dry as... a sheet of paper in Moses’ bible; item E4,
participant three)
Uncorrected proofs - John Benjamins Publishing Company
Chapter 15. Metaphoric competence in language
b. e topic which was typically associated with the physical world was projected on
to a psychological state or the reverse, and the resulting metaphor was not a familiar
English or French saying. (For example: We could tell by the look on the teacher’s face
that his anger was... like a rocket searching its target; item E4, participant forty)
c. e topic remained in its customary domain (sense modality or physical reference)
but a radical shi in perspective was required, and the resulting metaphor was not a
familiar English or French saying. (For example: When I was a child, I was frightened
of my grandma’s teeth soaking in the glass in the bathroom. ey made me think of...
an old wreck forgotten in the sea; item E8, participant twenty seven).
So long as the underlying rationale for the metaphor was an established or familiar one,
inclusion of a novel word or phrase did not qualify it as “novel”. Endings were scored
as “conventional metaphors” if the response was a familiar English saying. Endings were
scored as “literal” if the adjective remained in its customary domain. Endings were scored
as “inappropriate” if neither of the two judges could nd a meaning. Examples of items
that received each of these scores can be found in Appendix 5.
Scoring of the English part of the test was carried out by two independent native Eng-
lish speakers, and scoring of the French part of the test was carried out by two independent
native French speakers and one English speaker. All of these judges had been instructed
carefully in the scoring criteria. e English speaking judges achieved 98% agreement in
their assessment of responses, and the French speaking judges achieved 95% agreement.
In cases of disagreement, negotiations took place until agreement was reached. Most of
the replies given were conventional metaphors. Students also provided novel metaphors
and literal responses. ere were very few inappropriate responses. ese ndings match
those of both Trosborg (op. cit.) and Gardner et al. (op. cit.).
e sum of the scores was calculated for each student. A high score was interpreted
as meaning that the student had a preference towards metaphorical production, and a low
score was interpreted as meaning that a student had a preference towards literal produc-
tion. e reliability of nal version of the English version of this test (as measured by
Cronbach’s alpha) was 0.53 and the reliability of the French version was 0.58. Although
both of these reliabilities are fairly low, it was decided that, as no students would be af-
fected by any decisions made on the basis of their results in these tests, the tests would be
acceptable for research purposes.
. Results
It will be remembered that the research questions investigated in this study were as
follows:
1. If a student displays a high level of metaphoric competence in the L1, will they also
display a high level of metaphoric competence in the L2?
2. Are levels of L1 metaphoric competence higher or lower than levels of L2 metaphoric
competence?
Uncorrected proofs - John Benjamins Publishing Company
Jeannette Littlemore
. Findings for Research Question 1
e rst question was whether the scores on the metaphor tests in the L1 would be related
to the scores in the L2. is was found to be the case for all four tests. Scores on all four
tests in the L1 correlated with scores on the equivalent test in the L2. e results are pre-
sented below in Table 1:
Table 1. Correlations between English and French versions of the metaphor tests
Novel metaphor
production
(French)
French
metaphoric
uency
French
metaphor
interpretation
Speed of
French metaphor
interpretation
Novel metaphor
production (English)
0.33
p < 0.01*
−0.00
NS
0.12
NS
0.23
NS
English metaphoric
uency
0.18
NS
0.35
p < 0.01*
0.22
NS
0.12
NS
English metaphor
interpretation
0.19
NS
−0.04
NS
0.57
p < 0.01*
0.30
p < 0.01*
Speed of English meta-
phor interpretation
0.26
NS
0.21
NS
0.11
NS
0.59
p < 0.01*
is means that students who performed well on the tests in their rst language were sig-
nicantly more likely to perform well on the equivalent tests in their second language,
which indicates that a cognitive component is involved in metaphoric competence, as well
as a linguistic one. ese results also tell us that the English and French versions of the
metaphor tests display good convergent validity. e fact that only one other relationship in
the Table is signicant at the 99% level indicates that the discriminant validity is good too.
Within the languages themselves there were more relationships between the dierent
types of metaphoric competence. e correlations between the dierent tests in English
are shown in Table 2 below:
Table 2. Correlations between the dierent tests in English
Novel metaphor
production
(English)
English
metaphoric
uency
English
metaphor
interpretation
Speed of
English metaphor
comprehension
Novel metaphor
production (English)
0.18
NS
0.62
p < 0.01
−0.26
p < 0.05
English metaphoric
uency
0.04
NS
−0.16
NS
English metaphor
interpretation
−0.19
NS
Speed of English meta-
phor interpretation
e correlations between the dierent tests in French are shown in Table 3 below:
Uncorrected proofs - John Benjamins Publishing Company
Chapter 15. Metaphoric competence in language
Table 3. Correlations between the dierent tests in English
Novel metaphor
production
(French)
French
metaphoric
uency
French
metaphor
interpretation
Speed of
French metaphor
interpretation
Novel metaphor
production (French)
−0.19
NS
0.75
p < 0.01
−0.34
p < 0.01
French metaphoric
uency
−0.12
NS
−0.09
NS
French metaphor
interpretation
−0.31
p < 0.01
Speed of French meta-
phor interpretation
In both English and French, it seems that novel metaphor production has most in com-
mon with the other types of metaphoric competence. Because of these ndings we cannot
say that the dierent types of metaphoric competence are independent of one another.
is is perhaps to be expected as to some extent they reect the same cognitive processes
(see Littlemore and Low, 2006b, chapter 3).
. Findings for Research Question 2
e fact that the English and French versions of the metaphor tests correlated with each
other does not mean, however, that they were identical. e second research question was
interested in the relative levels of performance on the tests in the two languages. As we can
see in Table 4, the participants scored more highly in their rst language than in their sec-
ond language on the test of metaphoric uency, but in all the three other tests they scored
more highly in their second language than in their rst language:
Table 4. L1 and L2 comparisons of scores on all four tests
Language Novel
metaphor
production
Metaphoric
uency
Metaphor
interpretation
Speed of metaphor
interpretation (a low score
indicates a better result)
First language (French) 2.79 1.74 2.91 6.595
Second language (English) 2.83 1.36 3.03 6.478
In order to assess the signicance of these dierences, t-tests were carried out on all four
tests. ese tests revealed that for some of the tests there were signicant dierences in
level of performance between the two languages.
No signicant dierence was identied in the level of performance on the English and
French versions of the test of novel metaphor production t(81) = 1.11, p = 0.27). e level
of linguistic competence appears to have no eect on this type of metaphoric competence.
is nding appears to support Johnson’s (1989) contention that metaphoric competence
is a cognitive rather than a linguistic phenomenon.
Uncorrected proofs - John Benjamins Publishing Company
Jeannette Littlemore
ere was a signicant dierence in performance between the English and French
versions of the test of metaphoric uency t(81) = −5.65, p = 0.00). e results suggest that
the participants displayed higher levels of metaphoric uency in the L1 than in the L2.
is nding is probably due to shortcomings in the students’ L2vocabulary.
A signicant dierence was found between the English and French versions of the test
of metaphor interpretation t(81) = 2.34, p = 0.02). e results suggest that participants are
more likely to nd meaning in L2 metaphors than in L1 metaphors. is nding is dicult
to explain. It could be that in their L1 students have enough condence to claim that a
given metaphor makes no sense, whereasin their L2 students are used to having to guess
meanings and are therefore reluctant to dismiss expressions as unmeaningful.
ere was no signicant dierence between time taken to nd meaning in English
metaphors and time taken to do so in French metaphors t(81) = −0.30, p = 0.77). us
the only two signicant ndings were that: students were signicantly more likely to nd
meaning in metaphor in their second language, and that they exhibited signicantly high-
er levels of metaphoric uency in their rst language.
e nding that students performed worse in their L1 than in their L2 on the ten-
dency to nd meaning in metaphor test, although unusual, is in keeping with Bromberek-
Dyzman and Ewert’s nding (this volume) that Polish learners of English found it easier
to understand metaphor-based implicatures in English than in Polish. It is easier to un-
derstand the nding that students performed better in their L1 than in their L2 on the
metaphoric uency test. One would expect students to have greater access to the relevant
encyclopaedic knowledge in the L1 to help them perform this kind of test.
One reason for the dierences in performance patterns on these two tests may relate
to the fact that the tendency to nd meaning in metaphor test only required the students
to identify one meaning for the items concerned, whereas the metaphoric uency test
required them to nd multiple meanings. It could be that students are used to thinking
metaphorically in order to nd meaning in the L2 (Picken, 2007; Underwood et al., 2004),
whereas in their L1 so much of the language is processed formulaically that they no longer
need to practise this skill (Wray, 1999). On the other hand, metaphoric uency involves
deeper processing and the activation of a wider area of encyclopaedic knowledge within
the mental lexicon. As such, it is understandable that students will be better placed to do
this in their L1 than in their L2.
. Discussion
e ndings made in the rst part of this study suggest that the ability to understand
and produce metaphor in the L1 is related to the ability in the L2. In other words, stu-
dents who are able to do this in the L1 are also likely to do so in their L2. is suggests
that metaphoric competence is a relatively stable individual dierence variable, which
may partially account for dierences in student behaviour and success rates in the for-
eign language classroom. Further research in this area could usefully investigate the re-
lationship between dierent types of metaphoric competence and cognitive exibility.
Successful second language acquisition must involve a degree of cognitive exibility and
Uncorrected proofs - John Benjamins Publishing Company
Chapter 15. Metaphoric competence in language
openness to new ways of seeing things. Cognitive exibility is also likely to be related to
what Grigorenko et al. (2000) refer to as ‘Cognitive Ability for Novelty in Acquisition of
Language’. Basically, this involves the ability to spot new patterns in the language input and
to use one’s existing knowledge selectively, along with analogical reasoning, to work out
the new form-meaning pairing.
ese ndings underline the fact that metaphoric competence must not be seen as a
homogeneous trait. Rather, it is, to some extent, a multifaceted entity, and a student can,
for example, be good at nding the meaning in metaphor fairly quickly, but not good at
producing multiple meanings. Student performances on these tests were uncorrelated in
both English (see Table 2) and French (Table 3). In order to truly understand the ndings
made in this part of the study, more work is needed investigating the nature of the L1 and
L2 mental lexicons and the ways in which they overlap, and searches are carried out within
them. A second promising approach would be to explore the ways in which the acquisi-
tion of a second or third language extends and enriches the number of possible ways of
perceiving, describing and structuring our realities. is corresponds to Cook’s (2002)
notion of multicompetence, whereby linguistic knowledge is restructured in the mind of
a bilingual leading to an integrated system which combines elements from both the L1
and the L2 to produce something new. is integrated system is likely to aect the ways in
which learners understand and interpret metaphor.
e second part of the study indicates that students exhibit higher levels of meta-
phoric uency in their L1 than in the L2 but that they are more likely to nd meaning in
L2 metaphors than L1 metaphors. ere were no dierences in terms of performance on
the L1 and L2 versions of the tests of novel metaphor production and speed in nding
meaning. e rst of these ndings is perhaps to be expected. As we discussed above, this
test is likely to reect vocabulary knowledge in the two languages. e other ndings are
more surprising. e nding that students are more likely to nd meaning in L2 meta-
phors than L1 metaphors appears, at rst sight, to be somewhat counter-intuitive. One
possible explanation for this nding however is the fact that, as we saw above, the ratings
given by the students for the individual items correlated with ease of comprehension on
the English version of the test, whereas they correlated with goodness of metaphor on
the French version of the test. is indicates that they were looking for slightly dierent
things in these dierent parts of the test, which could explain the discrepancies in the re-
sults. Items that are being assessed for ease of comprehension will likely to receive higher
scores than items that are being assessed for ‘goodness’. is needs to be acknowledged as
a limitation of the study.
e study has a number of other limitations that could be remedied in future re-
search. For example, it has not really unpacked the issue of metaphoric competence versus
preference. Although the rubrics encouraged the participants to try their hardest with the
metaphor activities it could be that some students simply felt more motivated than oth-
ers to do well in the tasks and this may explain the signicant correlations across the
two languages. On the other hand, even if this were the case, the discovery of individual
dierences in terms of metaphoric preference is also of interest as a substantial amount
of second language learning involves thinking metaphorically. If learners vary in terms
of how much they like to do so, then this could perhaps underlie some of the dierential
Uncorrected proofs - John Benjamins Publishing Company
Jeannette Littlemore
success rates that teachers experience. Motivation to learn a second language is inevitably
a complex issue (Dornyei, 2001) but at some level it is likely to involve a love of language,
and a willingness to ‘play’ with the target language (Cook, 2000). It would be very interest-
ing to investigate how metaphoric competence relates to a propensity for language play.
Finally, this study has focussed on the comprehension and production of decontex-
tualised metaphors in highly articial test sessions. is of course is very dierent from
the sort of thing that learners will need to do in the real world. Future studies should look
more at naturally-occurring discourse, or compare the metaphor use of learners engaged
in more naturalistic activities in their rst and second language.
Notes
. Care must be taken when interpreting these ndings as there is likely to be variation in the children’s
tendencies to report on all the similarities they observe. However, the change from literal to metaphoric
understanding has been widely observed across a number of dierent types of metaphor tests. Some of
these tests (for example the Metaphoric Triads Test – see Kogan, 1983) contain a highly reduced report-
ing element.
. Prior to this test, the participants were given training in how to use the mouse. is training consisted
of a scale on the screen fromone to ve (see Appendix 2). A number appeared in the top le hand corner
of the screen and the participant had to use the mouse to click on the corresponding number in the scale.
If a number appeared that was not in the scale the participants were instructed to click on a box labelled
“other”. As soon as the participant had clicked in the appropriate place, another number appeared in the
top le hand corner of the screen and the participant had to repeat the procedure. is procedure con-
tinued until the average reaction time of the participant was below three seconds or until thirty trials had
taken place, whichever happened rst.
References
Billow, Richard (1975). A cognitive developmental study of metaphor comprehension. Developmental Psy-
chology 11, 415–423.
Bottini G., Corcoran R., Sterzi R., Schenone P., Scarpa, P., Frackowiak, R. S. J., & Frith, C. D. (1994). e
role of the right hemisphere in the interpretation of Figurative Aspects of Language: A Positron
Emission Tomography Activation Study. Brain 117 (2), 1241–125.
Bromberek-Dyzman, Katarzyna & Ewert, Anna. (is volume). Figurative competence is better devel-
oped in L1 than in L2, or is it? Understanding conversational implicatures in L1 and L2.
Cameron, Lynne (2003). Metaphor in educational discourse. London: Continuum.
Cometa, Michael, & Eson, Morris (1978). Logical operations and metaphor interpretation: A Piagetian
interpretation. Developmental Psychology 49, 649–659.
Cook, Guy (2000). Language play, language learning. Oxford: Oxford University Press.
Cook, Vivian (2002). Background to the L2 user perspective. In V. J. Cook (Ed.), Portraits of the L2 User
(1–32). Clevedon: Multilingual Matters.
Danesi, Marcel (2008). Conceptual errors in second language learning. In Sabine De Knop & Teun De
Rycker (Eds.), Cognitive approaches to pedagogical grammar (231–257). Berlin/New York: Mouton
de Gruyter.
Uncorrected proofs - John Benjamins Publishing Company
Chapter 15. Metaphoric competence in language
Dörnyei, Zoltan (2001). Teaching and researching motivation. Harlow: Longman.
Elbers, Lockie (1988). New names from old words: Related aspects of children’s metaphors and word
compounds. Journal of Child Language 15, 591–617.
Evans, Vyvyan & Green, Melanie (2006). Cognitive linguistics. An introduction. Edinburgh: Edinburgh
University Press.
Fauconnier, Gilles & Turner, Mark (1998). Conceptual integration networks. Cognitive Science 22 (2),
133–187.
Gardner, Howard, Kircher, Mary, Winner, Ellen & Perkins, David (1974). Children’s metaphoric produc-
tions and preferences. Journal of Child Language 2, 125–141.
Gibbs, Raymond (1994). e poetics of mind. Cambridge: Cambridge University Press.
Giora, Rachel (2003). On our mind. Salience, context, and gurative language. Oxford: Oxford University
Press.
Glucksberg, Sam, Newsome, M. R. & Goldvarg, Y. (2001). Inhibition of the literal: Filtering metaphor-
irrelevant information during metaphor comprehension. Metaphor and Symbol 16 (3 & 4), 277–293.
Gregory, Monica, & Mergler, Nancy (1990). Metaphor comprehension: In search of literal truth, possible
sense, and metaphoricity. Metaphor and Symbolic Activity 5/3, 151–173.
Grice, Howard(1975). Logic and conversation. In P. Cole & J. L. Morgan (Eds.), Syntax and Semantics3.
New York: Academic Press.
Grigorenko, E., Sternberg, R. & Ehrman, M. (2000). A theory-based approach to the measurement of
foreign language learning ability: the canal-f theory and test. e Modern Language Journal 84 (3),
390–405.
Holme, Randal (2004). Mind, metaphor and language teaching. Basingstoke: Palgrave Macmillan.
Johnson, Janice (1989). Factors related to cross-language transfer and metaphor interpretation in bilingual
children. Applied Psycholinguistics 10, 157–177.
Johnson, Janice (1991). Developmental versus language-based factors in metaphor interpretation. Journal
of Educational Psychology 83/4, 470–483.
Johnson, Janice (1996). Metaphor interpretations by second language learners: Children and adults. e
Canadian Modern Language review 53/1, 219–241.
Johnson, Janice & Rosano, Teresa (1993). Relation of cognitive style to metaphor interpretation and sec-
ond language prociency. Applied Psycholinguistics 14, 159–175.
Katz, Albert, Paivio, Alan, Marschark, Mark & Clark, James (1988). Norms for 204 Literary and 260 non-
literary metaphors on 10 psychological dimensions. Metaphor and Symbolic Activity 3/4, 191–214.
Kogan, Nathan (1983). Stylistic variation in childhood and adolescence: Creativity, metaphor, and cogni-
tive styles. In J. H. Flavell & E. M. Markman (Eds.), A handbook of child psychology3/4 (695–706).
New York: John Wiley and Sons.
Lako, George & Johnson, Mark (1980). Metaphors we live by. Chicago: University of Chicago Press.
Li, omas (2002). e acquisition of metaphorical expressions, idioms, and proverbs by Chinese learners of
English: A conceptual metaphor and image schema-based approach. PhD thesis (Chinese University
of Hong Kong).
Littlemore, Jeannette, & Low, Graham (2006a). Metaphoric competence and communicative language
ability. Applied Linguistics 27 (2), 268–294.
Littlemore, Jeannette, & Low, Graham (2006b). Figurative thinking and foreign language learning. Basing-
stoke: Palgrave Macmillan.
Ortony, Andrew (1979). Beyond literal similarity. Psychological Review 86, 161–180.
Ortony, Andrew (1980). Some psycholinguistic aspects of metaphor. In R. P. Honeck and R. P. Homan
(Eds.), Cognition and gurative language (69–83). New Jersey: Erlbaum Press.
Paivio, Alan & Walsh, M. (1993). Psychological orocesses in metaphor comprehension and memory, In
A. Ortony (Ed.), Metaphor and thought (307–308). 2nd ed. Cambridge: Cambridge University Press.
Pascual-Leone, J., & Burtis, J.(1975). Manual for FIT: Figural intersections test.Unpublished manuscript,
York University Department of Psychology, Toronto. Cited in Johnson J. (1991). Developmental
Uncorrected proofs - John Benjamins Publishing Company
Jeannette Littlemore
versus language-based factors in metaphor interpretation. Journal of Educational Psychology 83/4,
470–483.
Picken, Jonathan (2007). Literature, metaphor and the foreign language learner. Basingstoke: Palgrave
MacMillan.
Pollio, Howard & Smith, Michael (1979). Sense and nonsense in thinking about anomaly and metaphor.
Bulletin of the Psychonomic Society 13 (5), 323–326.
Trosborg, Anna (1985). Metaphoric productions and preferences in second language learners. In William
Paprotté and René Dirven (Eds.), e ubiquity of metaphor (525–557). Amsterdam: John Benjamins
Publishing Company.
Underwood, Georey, Schmitt, Norbert & Galpin, Adam (2004). e Eyes Have It: An eye-movement
study into the processing of formulaic Sequences. In N. Schmitt (Ed.), Formulaic sequences. (Amster-
dam/Philadelphia: John Benjamins Publishing Company).
Waggoner, James & Palermo, Diane (1989). Betty is a bouncing bubble: Children’s comprehension of
emotion-descriptive metaphors. Developmental Psychology 25/1, 152–163.
Woodcock, Richard (1980). Woodcock language prociency battery, English form. Hingham Mass.: Teach-
ing Resources.
Wray, Alison (1999). Formulaic language in learners and native Speakers. Language Teaching 32, 213–231
Appendix 1
English translation of the rubric used in the computer-based metaphor test
A metaphor is a statement which is not literally correct, but which establishes a relationship between two
parts of a sentence. e ease with which this relationship can be interpreted can vary.
For example, the statement “snow is a Winter coat” is an obvious metaphor – snow is not a coat, but the
idea of a coat provides relevant information about snow – it covers everything, it keeps you warm, it’s
thick...
On the other hand, the sentence “the piano is a spoon” cannot really be considered to be a metaphor: it is
dicult to see what kind of information a spoon can give about a piano. is kind of expression is known
as an anomaly.
You are going to be presented with two sets of sentences, in French and in English.
Your task is to decide to what extent each sentence can be said to be a metaphor, in other words, to what
extent one element can be said to provide information about the other.
You are not being asked to judge the quality of the metaphors, but simply to decide whether you think
they are metaphors or not.
Please indicate your response as soon as you have decided whether or not there is a relationship between
the two elements.
Uncorrected proofs - John Benjamins Publishing Company
Chapter 15. Metaphoric competence in language
Appendix 2
Metaphors used in the computer-based metaphor comprehension test
A dog is a walking stick
A dog’s stomach is his master’s alarm clock
A dream is a solar eclipse
A photograph is a one-sided skin of truth
A smile is a knife
A so-boiled egg is a guillotined aristocrat
A sugar-cube melting in coee is the fading of a ghost
Aeroplanes are angry birds
Beaches are grills
Creativity is a toaster
Death is a fat y
Death is the cruel singing of deathless mosquitoes
Evolution is a lottery
History is a sponge
Memory is a snake
Mimes are wooden statues
Music is death
Nature is a vast laboratory
Night is a castle
Smiles are the channels of future tears
e sky is a parliament
e wind is a cat
Time is a bird
Winter is the warm south
Wisdom is a weatherman
L’esprit createur est une bouilloire sur le feu
L’esprit est une eponge
L’histoire est un sport d’hiver
L’hiver est un oiseau avec des dents
L’homme est un collier qui cherche un chien
Uncorrected proofs - John Benjamins Publishing Company
Jeannette Littlemore
L’humiliation est un rideau
La liberté est un deuxième soleil
La lune est un chat dansant
La mort est un joueur de tambour
La parole est la semence de la misère
La sagesse est un météorologiste
La terre est une bouche parfumée
Le ciel est un ordinateur
Le clair de la lune est de l’eau de javel
Le subconscient est une arène
Les étoiles errantes sont des enfants qui ne connaîssent pas leur arithmétique
Les nuages sont des mondes en laine
Les panneaux d’achage sont des verrues sur le paysage
Un éléphant est une petite serre
Un chirurgien est un décorateur d’intérieurs
Un désir est un arc-en-ciel
Une île est un bouchon
Une armée est une mer vivante
Une dynastie est une pièce de théâtre
Une station d’essence est une oasis
Appendix 3
Answer sheet for metaphoric uency test
METAPHOR INTERPRETATION SHEET
NAME:
Use this booklet to write down all the relationships which you can see between the two elements of each
metaphor. Use English for metaphors 1 to 5 and French for metaphors 6 to 10.
Remember, the metaphors are:
1. A dog is a walking stick.
2. Nature is a vast laboratory.
3. Smiles are the channels of future tears.
4. A dream is a solar eclipse.
Uncorrected proofs - John Benjamins Publishing Company
Chapter 15. Metaphoric competence in language
5. Evolution is a lottery.
6. L’esprit créateur est une bouilloire sur le feu.
7. La liberté est un deuxième soleil
8. Les panneaux d’achage sont des verrues sur le paysage
9. Une armée est une mer vivante
10. Une station d’essence est une oasis
Please ask a teacher if you do not know the meaning of any of these words.
Appendix 4
Sentence starters used in the test of original metaphor production
English sentence starters
1. In Winter, the weather in Scotland is extremely cold. As soon as you go out of the house your face
starts to feel........
2. Tom hasn’t cleaned his room for ages and it’s starting to smell. e smell reminds me of..........
3. We could tell by the look on the teacher’s face that his anger was.......
4. e lake was a shining ................................................................................ at the bottom of the valley.
5. Peter’s violin playing isn’t marvellous, but compared to that of Alf it sounds like........
6. Dr. Livingstone had been walking across the Sahara for ve days without any water. His throat was
beginning to feel as dry as..........
7. What a beautiful day! e clear sky reminds me of........
8. When I was a child, I was always frightened of my grandma’s teeth soaking in the glass in the bath-
room. ey made me think of........
French sentence starters
1. Pierre était dans la salle. Il était évident qu’il avait vu quelque chose d’erayant. Il tremblait comme.........
2. Agnès renverse tout le temps quelque chose. On dirait.......
3. Les deux garçons entraient dans la hutte. Ils avaient très peur car il y faisait noir.......
4. Après sa maladie, mon père est devenu sourd........
5. La police l’avait presque rattrapé. Il se sentait.......
6. La sorcière était très âgée ; elle semblait presque avoir la peau.....
7. Il faut faire attention si tu sors avec ce mec là. J’ai entendu dire qu’il boît......
8. C’est vrai que maintenant elle est vieille et laide, mais quand elle était jeune, elle avait une peau........
Uncorrected proofs - John Benjamins Publishing Company
Jeannette Littlemore
Appendix 5
Examples showing how the items were scored in the metaphor production test
English Examples
e following were scored as novel metaphors:
We could tell by the look on the teacher’s face that his anger was... as terrible as the exam questions he gave
us last year. (item E3, participant 30)
e lake was a shining... and warm bath for the most terrible leviathan I had ever seen..., at the bottom of
the valley. (item E4, participant 20)
e following were scored as conventional metaphors:
Dr Livingstone had been walking across the Sahara for ve days without any water. His throat was begin-
ning to feel as dry as... the sand. (item E6, participant 54)
e lake was a shining... diamond ... at the bottom of the valley. (item E4, participant 58)
e following were scored as literal responses:
In winter the weather in Scotland is extremely cold. As soon as you go out of the house your face starts to
feel the cold winds from the North. (item E1, participant 1)
What a beautiful day! e clear sky reminds me of... my home country. (item E7, participant 33)
e following were scored as inappropriate responses:
When I was a child, I was always frightened of my grandma’s teeth soaking in the glass in the bathroom.
ey made me think of... old chewing gums. (item E8, participant 29)
e lake was a shining... river in fact... at the bottom of the valley. item E4, participant 62)
French Examples
e following were scored as novel metaphors:
Agnès renverse tout le temps quelque chôse. On dirait... qu’elle a été une boule de bowling dans une autre
vie. (item F2, participant 12)
(Agnes is always knocking things over. You might say... that in a previous life she was a ten pin bowling ball.)
La police l’avait presque rattrapé. Il se sentait coincé comme une praline Léonidas dans une boîte d’allu-
mettes. (item F5, participant 12)
(e police had almost caught him. He felt... as squashed as a Leonidas chocolate* inside a matchbox.)
* Leonidas chocolates are a particularly large kind of Belgian chocolates.
e following were scored as conventional metaphors:
C’est vrai que maintenant elle est vieille et laide, mais quand elle était jeune, elle avait une peau ... de
pêche*. (item F8, participant 19)
(It’s true that now she’s old and ugly, but when she was young she had skin... like a peach.*)
Uncorrected proofs - John Benjamins Publishing Company
Chapter 15. Metaphoric competence in language
Il faut faire attention si tu sors avec ce mec là. J’ai entendu dire qu’il boît... comme un trou.* (item F7,
participant 35)
(You’d better watch out if you go out with that bloke. I’ve heard it said that he drinks... like a hole.*)
*Conventional French idioms.
e following were scored as literal responses:
Agnès renverse tout le temps quelque chôse. On dirait...qu’elle fait exprès. (item F2, participant 1)
(Agnes is always knocking things over. You might say that... she does it on purpose.)
Les deux garçons entraient dans la hutte. Ils avaient très peur car il y faisait noir... et celle-ci se trouvait au
fond de la forêt. (item F3, participant 24)
(e two boys went into the hut. ey were very frightened because it was dark... and the hut was in the
depths of the forest.)
e following were scored as inappropriate responses:
Agnès renverse tout le temps quelque chôse. On dirait... qu’elle attire les objets. (item F2, participant 32)
(Agnes is always knocking things over. You might say... that she attracts objects.)
Agnès renverse tout le temps quelque chôse. On dirait... que c’est un peu pour elle. (item F2, participant 74)
(Agnes is always knocking things over. You might say... that it’s a bit for her.)
Uncorrected proofs - John Benjamins Publishing Company
