Content uploaded by Cristina Cacciari
Author content
All content in this area was uploaded by Cristina Cacciari on Jul 25, 2019
Content may be subject to copyright.
Processing ambiguous idioms
1
Accepted for publication in Journal of Cognitive Psychology
(DOI:10.1080/20445911.2015.1049178)
Literal analysis and idiom retrieval in ambiguous idioms processing: A reading-time study
Cristina Cacciari & Paola Corradini
Dipartimento di Scienze Biomediche, Metaboliche e Neuroscienze
Modena, Italy
Corresponding author:
Cristina Cacciari
Dipartimento di Scienze Biomediche, Metaboliche e Neuroscienze
Via Campi 287
41100 Modena, Italy
cacciari.cristina@unimore.it
Telephone: +39 059 2055339
Fax: +39 059 2055663
Running head: Processing ambiguous idioms
Processing ambiguous idioms
2
Processing ambiguous idioms
3
Abstract
We investigated the relationships between integration of the literal meaning of the idiom words and
activation of the idiomatic meaning using self-paced reading times. Participants read short stories in
which predictable and unpredictable ambiguous idioms were followed by literal or idiomatic
sentences disambiguating the contextually appropriate interpretation. Literal sentences were read
faster when preceded by unpredictable idioms than by predictable idioms, and faster than idiomatic
sentences when preceded by unpredictable idioms. Idiomatic sentences were read faster than literal
sentences when preceded by predictable idioms confirming that predictable idioms were resolved
idiomatically earlier on than unpredictable idioms. Idiomatic sentences were read equally fast no
matter whether they were preceded by predictable or unpredictable idioms suggesting that the
idiomatic meaning was available in both cases. Overall the results showed interplay between literal
compositional processes and idiomatic meaning retrieval modulated by the point at which the
idiomatic meaning became available.
Keywords: Idiom; Compositionality; Predictability; Comprehension; Context; Semantic ambiguity
Processing ambiguous idioms
4
Introduction
Generally, the meaning of a sentence is obtained by merging individual word meanings as
the message unfolds. However, sometimes what must be merged are unitary sequences of co-
occurring words, as for instance idioms, rather than single words. Idioms belong to the vast family
of multi-word sequences or expressions (MWEs) stored in semantic long-term memory (for
overviews, see Cacciari, 2014; Siyanova-Chanturia, 2013). MWEs are as frequent as words
(Jackendoff, 1995) representing an inescapable challenge to language processing models. The most
striking characteristic of idioms is that their meaning cannot be purely inferred from the constituent
parts since the relationship between lexical items and phrasal meaning is to a large extent arbitrary
and learned (we do not discuss here the contribution of the individual word meanings to the
figurative interpretation of idioms, for overviews see Cacciari, 1993; Cacciari, 2014). Although
idiomatic meaning is non-compositional, this does not mean that literal meanings play no role in
idiom processing. In fact, the words forming idiomatic configurations preserve their own lexico-
semantic properties while forming at the same time larger units (Molinaro et al., 2012).
But how can compositionality and non-compositionality go hand in hand? This seemingly
simple question has received different answers. Models of idiom processing in fact diverge on
whether or not the literal processing of the idiom words is carried out. For Lexical look-up models
(Swinney & Cutler, 1979; Bobrow & Bell, 1973; Gibbs, 1980), idioms are semantically empty long
words directly retrieved from the mental lexicon. Idiom recognition starts at the beginning of the
string and runs in parallel with the computation of literal meaning. Since computing takes longer
than retrieving, the idiomatic meaning becomes available before the literal meaning. Inconsistent
results questioned this view (e.g., Cacciari & Tabossi, 1988; Cacciari, et al., 2007; Peterson et al.,
2001; Titone & Connine, 1994; Tabossi et al., 2005; Holsinger & Kaiser, 2013; Libben & Titone,
2008; Vespignani et al., 2010) leading to the formulation of Non-Lexical Hybrid models (e.g.,
Cacciari & Tabossi, 1988; Titone & Connine, 1994; Sprenger et al., 2006). The most representative
version of this class of models is the Configuration Hypothesis (CH, Cacciari & Tabossi 1988) that
Processing ambiguous idioms
5
holds that idioms are incrementally processed word by word, just like any other piece of language,
until enough information has accumulated to signal the presence of an idiom. Only at this point the
idiomatic meaning is retrieved. Usually, the recognition point of predictable idioms is located in the
penultimate or antepenultimate word of the string (Cacciari & Tabossi, 1988; Tabossi et al., 2005;
Cacciari et al., 2007; Vespignani et al., 2010), and coincides with the last constituent of the idiom
string in unpredictable idioms.
Semantic ambiguity and idiomatic meaning
Some idioms have both an idiomatic and a well-formed, plausible literal interpretation (e.g.,
kick the bucket). Usually which of the two interpretations is appropriate is clarified by prior
linguistic context (Fanari et al., 2010). Although these ambiguous strings possess both an
idiomatically-plausible and a literally-plausible interpretation, in most cases the two interpretations
are not are equally frequent, and the figurative is dominant over the literal interpretation of the
idiom string (e.g., Popiel & McRae, 1988; Cronk & Schweigert 1992; Cronk, Lima & Schweigert,
1993; Cacciari et al., 2007). Are the literal and idiomatic meanings of ambiguous idioms processed
with equal ease or is there a processing priority of one interpretation over the other? Initial studies
(e.g., Gibbs, 1980; Mueller & Gibbs, 1987; Estill & Kemper, 1982) consistently showed that
idiomatic paraphrases were judged faster than literal paraphrases in idiomatic contexts. Recent
studies using real-time measures found less consistent results. Some studies in fact found activation
of the idiomatic meaning only in idiomatic contexts (Colombo, 1993, 1998). Other studies reported
activation of the figurative meaning in idiomatic contexts for predictable and unpredictable idioms
and only for predictable idioms in literal contexts (Cacciari et al., 2007). Recently, Holsinger and
Kaiser (2013) tested the comprehension of ambiguous phrasal verbs (e.g., look up, rush into)
preceded by contextual information disambiguating the intended meaning and followed by
sentences semantically consistent or inconsistent with this initial bias. The authors found no effects
of prior literal or idiomatic expectations when ambiguous phrasal verbs were followed by literal
Processing ambiguous idioms
6
sentences. Rather literal expectations slowed down the reading times of idiomatic sentences. When
readers expected a literal interpretation of the phrasal verb and instead the sentence turned out to be
idiomatic, processing was more disrupted than when readers were expecting an idiomatic
interpretation and the sentence turned out to be literal. In sum, the figurative interpretation of
ambiguous strings seems to have a temporal priority over the literal interpretation, if contextually
supported. At the same time evidence suggests that some literal analysis of the string is entertained.
Aims of the Present Study
This study investigated the interplay between integration of the literal meanings of the idiom
words and activation of the idiomatic meaning. Following the CH, we assume that this interplay is
dependent upon how early readers can predict the presence of an idiom in a sentence. Idiom
predictability corresponds to the probability of completing an idiom fragment idiomatically. The
CH was the first idiom comprehension model that explicitly assigned a central role to idiom
predictability. In fact, according to the CH, the notion of predictability captures the fact that, upon
reading part of a string, readers may recognize it as belonging to a known idiom. This in turn
triggers specific expectations toward the most likely continuation of the idiom fragment (the idiom-
remaining word(s)), as it has been shown also in recent behavioral and ERP studies (Vespignani et
al., 2010; Rommers et al., 2013; Molinaro & Carreiras, 2010; Molinaro et al., 2012). Usually idiom
predictability is empirically operationalized using cloze probability tests: subjects are asked to
complete idiom fragments of increasing length with the words that came to mind. The point at
which the fragment is completed idiomatically (originally called the idiom key in Cacciari &
Tabossi, 1988) indicates how early an idiom is recognized. Hence it allows to establish whether an
idiom is predictable or unpredictable before offset. For instance, while reading let the cat out of the
calls to mind bag, reading wet behind the typically elicits literal completions rather than the
idiomatic one (ears). The notion of idiom predictability has two facets. In principle, idiom
predictability develops in time as information accumulates, hence it can be considered a continuous
Processing ambiguous idioms
7
dimension. In fact, comprehenders’ propensity to complete an initial fragment of an idiom string
idiomatically depends on how much input is required for recognizing the idiomatic nature of the
string. However, in many studies (e.g., Cacciari et al., 2007; Titone & Connine, 1994; Tabossi &
Zardon, 1993, 1995; Tabossi et al., 2005) idiom predictability is transformed in a discrete,
dichotomous dimension (predictable vs. unpredictable) by setting a threshold that determines how
much of the input is necessary to prompt idiom recognition (1).
Although predicting what comes next in a sentence is a general characteristic of the human
sentence processing system (for an overview, see Kutas, DeLong, & Smith, 2011), expectations
concerning the expected word(s) in idioms (as in many other types of prefabricated unit) are
thought to be much more likely and precise than in freely occurring literal language since MWEs
are stored in semantic long-term memory in default form (Sprenger et al., 2006; Cacciari, 2014;
Molinaro, Barraze, & Carreiras, 2013). In fact, several studies (Vespignani et al., 2010; Molinaro &
Carreiras, 2010; Molinaro, Barraze, & Carreiras, 2013) showed that when the presence of an idiom
(or of a MWE) is identified before the string offset and the unit is activated, also the missing
constituent(s) are made available and compared with the pre-stored idiom representation in a
categorical identification process (Kok, 2001) aimed at monitoring whether or not it is the expected
item (Vespignani et al., 2010; Molinaro & Carreiras, 2010).
The CH holds that the idiomatic meaning becomes available when sufficient linguistic input
makes the configuration recognizable. However, the CH left rather unexplored the reverse side of
the story, namely whether and how the different amount of literal word meaning available before
idiom recognition affects the processing of predictable and unpredictable idioms. The meaning of
unpredictable idioms becomes available at the string offset (or shortly thereafter) when the
meanings of the individual idiom words are fully available. But far less literal meanings are
activated in predictable idioms since the idiomatic meaning is retrieved well before the string offset.
In this case, the literal meaning of the remaining idiom words (i.e., of the words following the
recognition point) may not be activated at all or may be rapidly suppressed (Peterson et al., 2001;
Processing ambiguous idioms
8
Rommers et al., 2013). As Rommers et al. (2013) argued, since the literal meaning of the idiom
words is generally unnecessary to compute the figurative meaning, the language system may not
engage, or may only partially engage, in semantically unifying individual word meanings while
processing idioms (where semantic unification refers to the integration of word meanings by
combining them into larger units, Hagoort, 2005) (2). However, how these processes are modulated
by predictability is not yet fully understood. Answering these questions, complementing unresolved
aspects of the CH, was the main aim of this study.
In this study, we used predictable and unpredictable ambiguous idioms followed by
sentences biasing toward the literal or idiomatic interpretation. Since it is well known that idiom
familiarity affects the time it takes to compute an idiomatic meaning (Schweigert, 1986; Cronk &
Schweigert 1992; Cronk, Lima & Schweigert, 1993; for overviews, see Libben & Titone, 2008;
Nordmann, Cleland & Bull, 2014), we only used highly familiar idiom strings. In predictable
idioms, the literal analysis may be switched off earlier than in unpredictable idioms making the
semantic integration of the figurative meaning in idiomatic sentences rather easy. But when the
upcoming sentence has a literal bias, the earlier commitment toward an idiomatic interpretation of
predictable idioms may have a processing cost due to the need to semantically reanalyze the idiom
string. Based on previous evidence (Cacciari et al., 2007; Titone & Connine, 1994), we assume that
the idiomatic meaning of unpredictable idioms and the constituent word literal meanings may be
both available at the idiom offset (or shortly thereafter). Hence, when unpredictable idioms are
followed by literal sentences, the semantic integration and unification of the literal meaning of the
idiom words can be easier with a rapid decay of the figurative interpretation that had less time to
consolidate and integrate in the sentence meaning. This should lead to faster reading times of literal
sentences than of idiomatic sentences when preceded by unpredictable idioms. In general, recovery
from misinterpretation should produce longer reading times of the disambiguating sentence, and
possibly also of the final region due to the possibility that readers may continue to process the
Processing ambiguous idioms
9
disambiguating sentence (spillover effect) while inspecting the final sentence to wrap-up a coherent
final interpretation to the story.
As in prior idiom studies (e.g., Colombo, 1993, 1998; Holsinger & Kaiser, 2013), we used a
self-paced reading time paradigm that has the double advantage that readers read at a pace that
corresponds to internal comprehension processes, and that no overt decision on idioms is required.
Method
Participants. Forty-eight students volunteered to participate in the study (26 females, mean
age 24.2, SD = 3.6, range 19-26 years). All were Italian native speakers with normal or corrected-
to-normal vision, unaware of the aim of the study. The research was carried out fulfilling ethical
requirements in accordance with the standard procedures of the University of Modena.
Materials.
Norming phase
Idiom familiarity and knowledge of the idiomatic meaning. Eighty idioms with both an
idiomatic and a literal interpretation (ambiguous idioms) were selected from a list of Italian idioms.
Thirty university students were asked to rate the familiarity of each idiom on a seven-point scale
(from 1: never heard to 7: heard very often) reported under each idiom. At the end of the
questionnaire, participants were asked to go back to the beginning and paraphrase each idiom. Since
the task was rather demanding in time, the eighty idioms were split into two questionnaires with
forty idioms each. We selected sixty-four idioms that were both highly familiar (M = 5.68, SD =
.98) and correctly paraphrased by 95% of the participants.
Age of Acquisition (AoA). This measure has scarcely been considered in idiom studies.
However, descriptive studies (Tabossi, Arduino, & Fanari, 2011; Bonin, Méot, & Bugaiska, 2013)
showed that the estimated age of acquisition of an idiom correlates with the familiarity and
knowledge of the idiomatic meaning, and is a reliable predictor of paraphrase verification times
(Bonin, Méot, & Bugaiska, 2013). Therefore we included this variable in the norming phase and
Processing ambiguous idioms
10
selected predictable and unpredictable idioms with similar AoAs (see Table 1). The AoA of idioms
was evaluated using a questionnaire listing the 64 idioms with a 7-point rating scale reported under
each idiom (1: 0-2 years, 2: 3-4 years, 3: 5-6 years, 4: 7-8 years, 5: 9-10 years, 6: 11-12 years, 7:
more than 13 years). Twenty-five students were requested to estimate the age at which they had
learned each idiom. Fifty-six idioms with a mean score ranging from 4 and 6 point were selected.
These idioms were learned, on average, between 9 and 11 years (M = 5.29; SD = .51), in line with
the ages reported in the idiom acquisition literature (Levorato & Cacciari, 1999).
Idiom recognition point. In order to assess the point at which the idiomatic nature of each
string was identified, a questionnaire was prepared with twenty-eight minimal contexts containing
the idiom without its last constituent (e.g., the boy broke the …) and fifty-six filler fragments of
similar length and structure. Twenty-four students were asked to complete each sentence with the
first word that came to mind. The idioms that were completed idiomatically with a cloze probability
of at least 70% were considered as predictable and those that received less than 30% of idiomatic
completions as unpredictable. To ensure that the idiomatic completions were not due to literal
interpretations, each participant was also asked to provide a meaning paraphrase of each sentence.
Participants always provided idiomatic paraphrases. A total of 32 idioms were selected, 16
predictable and 16 unpredictable (cloze values of 79.9% and 13.9%, respectively, see Table 1).
Predictable and unpredictable idioms included an equal number of long and short strings.
Semantic transparency. To assess whether the literal and idiomatic meaning of the idiom
string bear any semantically transparent relationship, twenty-five participants were asked to rate the
extent to which the words forming each of the 32 idiom strings contributed to the figurative
interpretation (1: the words do not contribute at all to 5: the words contribute to the figurative
meaning). The idiomatic meaning of predictable and unpredictable idioms was equally mildly
transparent (see Table 1).
Meaning Dominance. To test whether the dominant meaning of idiom strings was literal or
figurative, twenty-five participants were asked to rate the extent to which each idiom strings was
Processing ambiguous idioms
11
heard/used in a literal or figurative sense (1: mostly literally to 5: mostly figuratively). The
figurative meaning was similarly perceived as the dominant one in predictable and unpredictable
idioms (Table 1).
Story plausibility. To test whether the idiomatic and literal stories associated to each idiom
were perceived as equally plausible, two lists were created containing either the idiomatic or the
literal versions of each story. 48 participants (24 for each list) were asked to rate the extent to which
each story was plausible (1: implausible to 7: totally plausible). The idiomatic and literal stories
associated to predictable and unpredictable idioms were perceived as equally plausible (Idiomatic
stories: M = 4.66, SD = .9; M = 4.96, SD = 1.1; Literal stories: M = 4.43, SD = .8; M = 4.83, SD =
.98; for predictable and unpredictable idioms, respectively; see Table 1).
The experiment used 32 ambiguous idioms, 16 predictable (PI) and 16 unpredictable (UP). The
idiom strings were highly familiar, had dominant figurative interpretations and moderately
semantically transparent meanings (Table 1). Idioms were embedded in short stories biased toward
the literal or the idiomatic meaning (see Table 2 for examples and the Appendix for the entire set of
experimental stories). A questionnaire containing the sixty-four stories in a randomized order was
submitted to ten students who judged whether each idiom string was meant literally or
idiomatically. Idiom strings were always interpreted literally in literal stories, and figuratively in
idiomatic stories.
Each story was formed by: 1) an Introductory Region composed by an idiom-neutral sentence
(mean number of characters = 18.9, SD = 3.1); 2) an Idiom Region where the idiom string was
preceded by an animate subject (number of characters of PI: M = 28.9, SD = 4.8; UP: M = 27.2, SD
= 3.7; t (30) = 1.115, SE 1.5, p = .27). The Introductory Region and the Idiom Region were identical
in the Idiomatic and Literal story of each idiom; 3) a Disambiguating Region biased toward the
literal or the idiomatic meaning of the string (PI: M = 33.5, SD = 2.4; UP: M = 33.6, SD = 1.8; t
(30) = .125, SE .75, p = .90; mean number of words of the literal version: 20.7, SD = 1.35;
idiomatic version: 20.5, SD = 1.3; t (31) = .961, SE .16, p = .34); 4) a Final Region completing the
Processing ambiguous idioms
12
stories in contextually coherent ways [PI: M = 18.6, SD = 2.5; UP: M = 19.6, SD = 2.6; t (30) =
1.184, SE .84, p = .25]. The literal version of the Disambiguating region did not contain any
constituents semantically associated to the idiom. Overall, the Disambiguating region was formed
by 28 final and 36 temporal subordinate sentences. In each idiom, the literal version and idiomatic
version of the Disambiguating region had the same type of subordinate sentence (either final or
temporal). We chose final and temporal sentences because of their frequency and relative ease of
processing (e.g., De Vincenzi & Job, 1995; Garrod et al., 1993). We balanced the cumulative
frequency of the words forming the literal and idiomatic Disambiguating regions of predictable and
unpredictable idioms (see Table 1).
Two lists were created so that each idiom was presented only once. Each list contained 184
stories in a quasi-randomized order: 64 experimental stories (32 stories forming the Literal
condition, 16 with predictable and 16 with unpredictable idioms; and 32 stories forming the
Idiomatic condition, 16 with predictable and 16 with unpredictable idioms) and 120 filler stories of
similar syntactic structure and length (mean number of words 20.6, SD = 1.4). Each list was divided
into two blocks of the same length separated by a short break. Sixteen of the filler stories,
interspersed among the experimental stories, ended with a double-choice question (True vs. False)
inserted for controlling whether participants read for comprehension. Half of the questions were
true and half false. Pressing the space bar after the last word of filler stories caused the item to be
replaced by a Yes/No comprehension question to which participants answered by pressing one of
two keys on the keyboard. A threshold of 75% of correct answers was used for a participant to be
included in statistical analyses. The experiment was preceded by a practice session formed by 10
different literal stories, divided in four regions, that were not further analyzed.
-------------------------------------------------------------------------------------------------------------
Table 1 around here
Processing ambiguous idioms
13
-------------------------------------------------------------------------------------------------------------
Procedure
The stimuli were presented on a computer screen using the DMDX2 software (Forster &
Forster, 2003). The computer recorded the time between each button-press as well as the responses
to the comprehension questions. Participants sat in a semi-dark sound-attenuated room at a distance
of approximately 65 cm from the computer screen. The instructions were presented on the screen
and were read by the experimenter. Participants were asked to read each sentence accurately for
comprehension and to press the spacebar with their dominant hand after reading each sentence to
make the next region appear. Each new story was preceded by a beep. The sentences appeared in
the center of the screen written in white characters (Geneva BOLD 14) on a black background.
Participants were informed that some true/false questions would appear during the experiment, and
that they had to respond by pressing one of two different keys that were marked with V (Vero,
True) or F (Falso, False). The experiment lasted approximately 30 minutes and was interrupted by
a short break after the first block. The order of the two blocks was counterbalanced. An equal
number of participants were assigned to each list.
Results
Two participants fell below the 75% threshold and were excluded from subsequent analyses.
The mean percentage of correct responses of the remaining 46 participants was 81%. To reduce
variability, data points +/- 2 SD from the mean reading times of each participant were excluded
from the analyses (5.2%). By-participant and by-item ANOVAs were conducted on the mean
reading times in each region using a 2 x 2 design with Story (Literal vs. Idiomatic) and Idiom
(Predictable vs. Unpredictable) as factors. In the by-participant analyses, Story and Idiom were
within-subject factors. In the by-item analyses, Idiom was a between-item factor and Story a within-
item factors. We analyzed the reading times of the Idiom Region, the Disambiguating Region and
Processing ambiguous idioms
14
the Final Region (Table 3). Since the mean number of characters of the different regions and
conditions were matched, ANOVAs were conducted on the mean reading times per region. By-item
Pearson partial correlations between each of the psycholinguistic variables (i.e., AoA, predictability,
semantic transparency, and meaning dominance) and the reading times of idiomatic and literal
conditions in the three regions were calculated partialling out the effects of all remaining ones (α =
.05; we only report statistically significant correlations). We present the reading times and partial
correlations relative to each region separately.
The Idiom Region
We collapsed the Literal and Idiomatic Story conditions since this region only contained the
idiom string preceded by an animate subject. Idiom was significant in the by-subject but not in the
by-item analysis [t1 (45) = 3.023, SEM = 18.42, p = .004; t2 (30) = .561, SEM = 74.43, p = .58] with
predictable idioms read faster than unpredictable idioms (by an average of 56 ms; 95% CI: + 37
ms). Partial correlations showed significant inverse correlations between AoA and response times
and between cloze values of predictable idioms and response times in that late-acquired idioms and
predictable idioms with higher cloze values were also read faster [r = -.433, p = .022; r = -.627, p =
.029; respectively].
-------------------------------------------------------------------------------------------------------------
Tables 2 and 3 around here
-------------------------------------------------------------------------------------------------------------
The Disambiguating Region
Processing ambiguous idioms
15
Crucially, we obtained a significant Story by Idiom interaction [F1 (1, 45) = 28.399, MSE =
14922.99, p = .0001, ηp2 = .387; F2 (1, 30) = 20.454, MSE = 10909, p = .0001, ηp2 = .405; MinF’
(1, 66) = 11.89, p = .001] (Figure 1). When preceded by predictable idioms, idiomatic continuations
were read faster than literal continuations (by an average of 112 ms; 95% CI: + 42.86 ms) [t1 (45) =
5.264, SEM = 21.28, p = .0001; t2 (15) = 3.212, SEM = 37.06, p = .006]. When preceded by
unpredictable idioms, literal continuations were read faster than idiomatic continuations (by an
average of 80 ms; 95% CI: + 54.23 ms) [t1 (45) = 2.968, SEM = 26.93, p = .005; t2 (15) = 3.184,
SEM = 36.79, p = .006]. Literal continuations were read faster when preceded by unpredictable
idioms than by predictable idioms (by an average of 150 ms; 95% CI: + 41.52 ms) [t1 (45) = 7.243,
SEM = 20.63, p = .0001; t2 (30) = 3.987, SEM = 42.18, p = .001]. Idiomatic continuations were
read slightly faster when preceded by predictable idioms than by unpredictable idioms (by an
average of 42 ms; 95% CI: + 42.56 ms) but the difference was not statistically significant (p1 = .14,
p2 = .21). Partial correlations confirmed the associations of high predictability values with slower
reading times of literal continuations [r = .589, p = .001] and faster reading times of idiomatic
continuations [r = -.394, p = .042].
-------------------------------------------------------------------------------------------------------------
Figure 1 around here
-------------------------------------------------------------------------------------------------------------
The Final Region
We obtained a main effect of Story, significant in the by-subject but not in the by-item
analysis [F1 (1, 45) = 4.091, MSE = 8321.951, p = .049, ηp2 = .083; F2 (1, 30) = .005, MSE =
Processing ambiguous idioms
16
9370.92, p = .46, ηp2 = .000], with the Final region read slightly slower in literal than in idiomatic
stories (by an average of 28 ms; 95% CI: + 25.93 ms). No significant partial correlations emerged.
Discussion and Conclusions
Our general aim was to determine whether and how compositionality and non-compositionality
went hand in hand in processing predictable and unpredictable ambiguous idioms embedded in
literal and idiomatic stories. We hypothesize that the interplay between integration of the literal
meanings of the idiom words and activation of the idiomatic meaning was dependent upon idiom
predictability. Before summarizing and discussing the main results, two limitations of this study
should be mentioned. First, since we compared the same idiom strings embedded in literal and
idiomatic stories, we did not have a baseline condition that would have allowed us to decide if
idiom activation speeded up the reading of idiom-resolving continuations or conversely literal
activation slowed down the reading of idiom-resolving continuations. Future studies are needed to
adjudicate on it. Second, due to the need of clearly orientating the reader toward the literal or
idiomatic interpretation of the idiom string, we could not use the same sentences in the literal and
idiomatic versions of the Disambiguating region of predictable and unpredictable idioms. However,
we tried to make the literal and idiomatic versions as psycholinguistically similar as possible by
matching them for semantic plausibility, number of words and characters, overall frequency of the
constituent words and type of subordinate sentence (See the Method section and Table 1).
The results showed that overall the reading times of the Idiom region were faster for predictable
idiom strings than for unpredictable ones (the association of faster reading times to high cloze
values was also confirmed by the partial correlation). Presumably this reflects the fact that reading
some of the words composing highly predictable idioms sufficed to actively pre-activate the
word(s) that were more likely to appear (Molinaro & Carreiras, 2010; Molinaro et al., 2013;
Processing ambiguous idioms
17
Vespignani et al., 2010). This is also consistent with prior evidence showing that predictable words
are read faster and more often skipped during natural reading (Erlich & Rayner, 1981).
Literal continuations were read faster when preceded by unpredictable idioms than by
predictable idioms, and faster than idiomatic continuations when preceded by unpredictable idioms.
This confirms that the literal interpretation of unpredictable idioms was left open (due to the fact the
literal meanings of the idiom word were fully available at the string offset) and was then easier to
integrate in literal than in idiomatic continuations. In fact, the availability of the literal word
meanings of unpredictable idioms advantaged readers (Holsinger & Kaiser, 2013) leading to a
smooth literal semantic unification of the idiom words (attested by the fastest reading times) when
the upcoming sentence was literal. The statistically equal reading times of idiomatic continuations
when preceded by predictable or unpredictable idioms can be taken to suggest activation of the
idiomatic meaning in both cases (consistently with the results of Cacciari et al., 2007; Titone &
Connine, 1994). How can we reconcile this result with the faster reading times of literal than
idiomatic continuations when preceded by unpredictable idioms? The recognition point of
unpredictable idioms was at the end of the idiom string, hence the idiomatic meaning started to
become available when the literal meaning of the idiom constituent words was fully active. The
literal bias of the upcoming sentence enhanced the literal interpretation. In contrast, the figurative
meaning of unpredictable idiom strings had less time to consolidate and could not integrate in the
sentential meaning. Therefore, it was rapidly suppressed being contextually incoherent. Idiomatic
continuations were read faster than literal continuations when preceded by predictable idioms
confirming the ease of integration of the dominant, figurative meaning (Holsinger & Kaiser, 2013)
in idiomatic sentences. This led to a rapid decay of the activated literal meaning of the few idiom
words preceding the idiom recognition point.
A striking result of this study is that the integration of predictable idioms in literal contexts was
extremely costly, as shown by the longest reading times. This result may reflect the fact that since
the idiomatic meaning of predictable idioms was activated well before the idiom offset, readers
Processing ambiguous idioms
18
strongly committed to the figurative interpretation. Activation of the literal meanings of the
remaining idiom constituent words (i.e., of the words following the idiom recognition point) was
irrelevant and may have been rapidly switched off (Rommers et al., 2013). When the sentence
following a predicable idiom turned out to be literal rather than idiomatic, readers had to revise their
interpretation and compute afresh the literal interpretation of the string. The processing cost of this
repair led to the highest reading times.
The present study was not meant to tease apart the predictions of lexical and non-lexical
models of idiom comprehension. Notwithstanding, while the results of this study can be easily
explained by the CH (we return on this point below), can be they as much easily accommodated by
other models of idiom comprehension? Lexical look-up models (e.g., Swinney & Cutler, 1979;
Gibbs, 1980) assume that the computation of the literal meaning of the idiom string and the retrieval
of the idiomatic meaning from the mental lexicon are separate and unrelated processes. Therefore,
these models assign no explicit role to idiom predictability not they assume that literal computation
can be part of the processing of the idiomatic meaning. Hence it is not clear what these models
could predict with respect to the questions tested in the present study. Turning to Non- lexical
models, aside from the CH, also the Superlemma Hypothesis (SH, Sprenger et al., 2006) may in
principle account for our results, provided that it further details how idiom comprehension proceeds
in time, and it specifies the role of predictability in comprehension. In fact, the SH was originally
conceived as a production model and is thought to operate top-down (from concepts to lemmas) in
production. As a comprehension model, the SH operates bottom-up in that the linguistic input
activates the single lemmas corresponding to the idiom constituent words. This in turn leads to the
activation of the corresponding conceptual nodes, and of the superlemma that triggers the activation
of the appropriate conceptual node. Activation of the superlemma starts from the first word and
increases as the perceptual input unfolds until a threshold is reached leading to idiom recognition.
As Sprenger et al. (2006) acknowledged, the SH is partly compatible with the CH since both
assume that literal computation is associated to idiom processing, and posit an activation threshold
Processing ambiguous idioms
19
for the idiomatic meaning. This threshold may be thought to be associated with idiom predictability,
in line with the CH. However, the CH and the SH differ in an important aspect concerning the point
in the string at which idiom activation starts. According to the CH, idiomatic meaning activation is
necessarily preceded by the recognition of the idiomatic nature of the input. This can be triggered
by part of the idiom string or by the entire idiom string (in unpredictable idioms). In contrast, for
the SH idiom activation builds up from the idiom string onset. From there onward, the cohort of
idioms starting with the same lexical entry should be activated, together with non-idiomatic
meaning(s). As we said, this study was not devised to tease apart these two hypotheses, and future
studies using similar experimental idioms and experimental techniques are needed to directly
compare them and decide whether indeed idiom activation starts from the idiom onset (being so
small at the beginning to get undetected), of only after the idiom recognition point. If indeed idiom
activation starts from the string first constituents, one might expect that the electrophysiological
signature of idiom processing would not change before and after the idiom recognition point. But
the picture emerging from some recent EEG studies on idioms and collocations revealed a different
picture (Vespignani et al. 2010; Molinaro & Carreiras, 2010; Molinaro et al., 2012; 2013; Siyanova-
Chanturia et al., 2011) showing that qualitatively different cognitive processes, indexed by different
electrophysiological components (N400 and P300, respectively), occurred before and after the
recognition of the idiomatic/collocational nature of the input. This may potentially weaken the
assumption that idiom activation is a linear process that starts from the activation of an idiom
cohort at the first idiom constituent. But again specific EEG studies directly comparing the
predictions of the CH and the SH are needed to adjudicate on this problem.
Idiom predictability refers to the fact that in some cases (predictable idioms) the presentation of
part of an idiom can prompt the recognition of the entire idiom string. It is likely that idiom
familiarity can affect the amount of input necessary to unveil the presence of an idiom. Indeed,
significant correlations between familiarity and predictability have been observed in descriptive
studies on idioms (Bonin, Méot, & Bugaiska, 2013; Libben & Titone, 2008; Tabossi, Arduino &
Processing ambiguous idioms
20
Fanari, 2011) and were found in the present study as well (rho = .36, p = .043). But, to the best of
our knowledge, so far no systematic investigations of the conjoint real-time effect of familiarity and
predictability have been undertaken (for an extensive discussion of idiom familiarity, see
Nordmann, Cleland & Bull, 2014). Future studies orthogonally manipulating these two factors are
needed to further clarify this issue. In addition, several studies (e.g., Popiel & McRae, 1988; Cronk
& Schweigert 1992; Cronk, Lima & Schweigert, 1993; Cacciari et al., 2007) showed that the two
interpretations of ambiguous idioms are rarely equally familiar and /or frequent and that usually
figurative meaning dominance facilitates the activation of the figurative interpretation. In our study,
the idiomatic interpretation can be thought to correspond to a “prepotent response” (Roberts, Hager
& Heron, 1994), namely, the most frequent response based on previous associations (e.g., frequency
information). A further interesting question concerns the relationship between meaning dominance
and literality. In most of the recent studies, the figurative interpretation was dominant over the
literal one, hence the time-course of figurative meaning activation in idioms where the literal and
idiomatic meaning are equally frequent is still poorly understood.
Overall, we showed that the literal meanings of the idioms words (or at least part of them) were
activated and that this could precede idiom retrieval depending on the idiom type (cf. Holsinger &
Kaiser, 2013; Canal et al., 2015). This suggests that the interplay between integration of the literal
meaning of the idiom words and the retrieval of the idiomatic meaning in ambiguous idioms is more
articulated than it was originally proposed by the CH in 1988. Indeed it seems that the
comprehension of ambiguous idioms involves a shift in the balance of literal and figurative analysis
that heavily relies on how early the idiom is detected. This further confirms that probabilistic
information about word co-occurrences is part of our knowledge of language; and that distributional
information is incorporated into the incremental processing of upcoming words (for an overview,
see Kutas et al., 2011). In fact the best way of specifying an idiom recognition point is in
probabilistic terms as the point after which the probability of a fragment to be continued
idiomatically is extremely high. This type of probability does not concern single words but rather
Processing ambiguous idioms
21
reflects the fact that certain words co-occur more frequently when part of MWEs than of freely
formed sentences (Vespignani et al., 2010). This is one of the reasons why the study of literal and
non-literal MWEs provides relevant insights about how the mind stores lexico-semantic knowledge
and process it, and about the ways in which semantic storage and processing interact with the
distributional properties of a language.
Processing ambiguous idioms
22
References
Bobrow, S., & Bell, S. (1973). On catching on to idiomatic expressions. Memory & Cognition, 1,
343-346.
Bonin, P., Méot, A., & Bugaiska, A. (2013). Norms and comprehension times for 305 French
idiomatic expressions. Behavior Research Methods, 45, 1259-1271.
Cacciari, C. (1993). The place of idioms in a literal and metaphorical world. In C. Cacciari, P.
Tabossi (Eds.). Idioms. Processing, Structure and Interpretation (pp. 27-55). Hillsdale, NJ:
Laurence Erlbaum
Cacciari, C. (2014). Processing multiword idiomatic strings: Many words in one? The Mental
Lexicon, 9, 2, 267-293.
Cacciari, C., & Tabossi, P. (1988). The comprehension of idioms. Journal of Memory and
Language, 27, 668-683.
Cacciari, C., Padovani, R., & Corradini, P. (2007). Exploring the relationship between individuals’
speed of processing and their comprehension of spoken idioms. European Journal of
Cognitive Psychology, 19, 3, 417-445.
Canal, P., Pesciarelli, F. Molinaro, N., Vespignani, F., & Cacciari, C. (2015). Electrophysiological
correlates of idiom comprehension: Semantic composition does not follow lexical retrieval. In V.
Pirrelli, C. Marzi & M. Ferro (Eds), Proceedings of the NetWordS Final Conference on Word
Knowledge and Word Usage: Representations and Processes in the Mental Lexicon, Pisa, Italy,
March 30 - April 1, 2015, pp. 98-101, http://ceur.org/vol. 1347.
Colombo, L. (1993). The comprehension of ambiguous idioms in context. In Cacciari, C. &
Tabossi, P. (Eds.), Idioms. Processing, Structure and Interpretation (pp. 3-26). Hillsdale, NJ:
Laurence Erlbaum.
Processing ambiguous idioms
23
Colombo, L. (1998). Role of context in the comprehension of ambiguous Italian idioms. In D.
Hillert (Ed.), Sentence processing: A cross-linguistic perspective. Syntax and Semantics, vol.
31. (pp. 405-425). New York: Academic Press.
Cronk, B. & Schweigert, W. A. (1992). The comprehension of idioms: The effects of familiarity,
literalness and usage. Applied Psycholinguistics, 13, 131-146.
Cronk, B. C., Lima, S. D., & Schweigert, W. A. (1993). Idioms in sentences: Effects of frequency,
literalness, and familiarity. Journal of Psycholinguistic Research, 22, 59–92.
Ehrlich, S. F., & Rayner, K. (1981). Contextual effects on word perception and eye-movements
during reading. Journal of Verbal Learning and Verbal Behavior, 20, 641-655.
Estill, R. B., & Kemper, S. (1982). Interpreting idioms. Journal of Psycholinguistic Research, 11,
559-568.
Fanari, R., Cacciari, C., & Tabossi, P. (2010). The role of idiom length and context in spoken idiom
recognition. European Journal of Cognitive Psychology, 22, 3, 321-334.
Forster, K.I., & Forster, J.C. (2003). DMDX: A windows display program with millisecond
accuracy. Behavior Research Methods, Instruments & Computers, 35, 116-124.
Gibbs, R. B. (1980). Spilling the beans on understanding and memory for idioms. Memory &
Cognition, 8, 449-456.
Glucksberg, S., Gildea, P., & Bookin, M.B. (1982). On understanding non-literal speech: Can
people ignore metaphors? Journal of Verbal Learning and Verbal Behavior, 21, 85-98.
Hagoort, P. (2005). On Broca, brain, and binding: A new framework. Trends in Cognitive Sciences,
9, 9, 416-423.
Holsinger E., & Kaiser E (2013). Processing (non)compositional expressions: mistakes and
recovery. Journal of Experimental Psychology: Learning, Memory and Cognition, 39, 3, 866-
878.
Kok, A. (2001). On the utility of P3 amplitude as a measure of processing capacity.
Psychophysiology, 38, 557-577.
Processing ambiguous idioms
24
Kutas, M., DeLong, K. A., & Smith, N. J. (2011). A look around at what lies ahead: prediction and
predictability in language processing. In M. Bar (Ed.), Predictions in the brain. Using our
past to generate a future (pp.190-207). New York: Oxford University Press.
Levorato, M. C., & Cacciari, C. (1999). Idioms comprehension in children: Are the effects of
semantic analyzability and context separable? European Journal of Cognitive Psychology, 11,
1, 51-66.
Libben, M., & Titone, D. (2008). The multi-determined nature of idiom processing. Memory &
Cognition, 36, 6, 1103–1121.
Molinaro, N., & Carreiras, M. (2010). Electrophysiological evidence of interaction between
contextual expectation and semantic integration during the processing of collocations.
Biological Psychology, 83, 3, 176–190.
Molinaro, N., Carreiras, M. & Duñabeitia, J.A. (2012) Semantic combinatorial processing of non-
anomalous expressions. Neuroimage, 59, 4: 3488-3501.
Molinaro, N., Barraza, P., & Carreiras, M. (2013). Long-range neural synchronization supports fast
and efficient reading: EEG correlates of processing expected words in sentences.
NeuroImage, 72, 120-132.
Mueller, R. A. G., & Gibbs, R. (1987). Processing idioms with multiple meanings. Journal of
Psycholinguistic Research, 16, 1, 63-81.
Nordmann, E., Cleland, A. A., & Bull, R. (2014). Familiarity breeds dissent: Reliability analyses for
British-English idioms on measures of familiarity, meaning, literality and decomposability.
Acta Psychologica, 140, 87-95.
Peterson, R. R., Burgess, C., Dell, G. S., & Eberhard, K.L. (2001). Dissociation between syntactic
and semantic processing during idiom comprehension. Journal of Experimental Psychology:
Learning, Memory and Cognition, 90, 227-234.
Processing ambiguous idioms
25
Pickering, M. J., & Traxler, M. J. (1998). Plausibility and recovery from garden-paths: An eye-
tracking study. Journal of Experimental Psychology: Learning, Memory and Cognition. 24,
4, 940-961.
Popiel, S.J. & McRae, K. (1988). The figurative and literal senses of idioms, or all idioms are not
used equally. Journal of Psycholinguistic Research, 17, 475-487.
Rommers, J., Dijkstra, T., & Bastiaansen, M. C. M. (2013). Context-dependent semantic processing
in the human brain: Evidence from idiom comprehension. Journal of Cognitive Neuroscience,
25, 5, 762-776.
Schweigert, W. A. (1986). The comprehension of familiar and less familiar idioms. Journal of
Psycholinguistic Research, 15, 33-45.
Roberts, R.J., Hager, L.D. & Heron, C. (1994). Prefrontal cognitive processes: Working memory
and inhibition in the antisaccade task. Journal of Experimental Psychology: General, 123, 374-
393.
Siyanova-Chanturia A., Conklin, K. & van Heuven, W. J. B. (2011). Seeing a phrase ‘time and
again’ matters. Journal of Experimental Psychology: Learning Memory and Cognition, 37, 3,
776-784.
Siyanova-Chanturia A. (2013). Eye-tracking and ERPs in multiword expression research. The
Mental Lexicon, 8, 2: 245-268.
Sprenger, S. A., Levelt, W. J., & Kempen, G. (2006). Lexical access during the production of
idiomatic phrases. Journal of Memory and Language, 54: 161-184.
Swinney, D., & Cutler, A. (1979). The access and processing of idiomatic expressions. Journal of
Verbal Learning and Verbal Behaviour, 18, 523-534.
Tabossi, P. & Zardon, F. (1993). The activation of idiomatic meaning in spoken language. In C.
Cacciari & P. Tabossi (Eds.). Idioms. Processing, Structure and Interpretation (pp. 145-162).
Hillsdale, NJ: Laurence Erlbaum.
Processing ambiguous idioms
26
Tabossi, P., & Zardon, F. (1995). The activation of idiomatic meaning. In M. Everaert, E. van den
Linden, A. Schenk & R. Schreuder (Eds.), Idioms: structural and psychological perspectives
(pp. 273-282). Hillsdale, NJ: Erlbaum.
Tabossi, P., Fanari, R., & Wolf, K. (2005). Spoken Idiom Recognition: Meaning Retrieval and
Word Expectancy. Journal of Psycholinguistic Research, 34, 465-495.
Tabossi, P., Arduino, L., & Fanari, R. (2011). Descriptive norms for 245 Italian idiomatic
expressions. Behavior Research Methods, 43, 110-123.
Titone, D.A., & Connine, C. N. (1994). Comprehension of idiomatic expressions: Effect of
predictability and literality. Journal of Experimental Psychology: Learning, Memory, and
Cognition, 20, 1126-1138.
Vespignani, F., Canal, P., Molinaro, N., Fonda, S., & Cacciari, C. (2010). Predictive mechanisms in
idiom comprehension. Journal of Cognitive Neuroscience, 22, 8, 1682-1700.
Processing ambiguous idioms
27
Footnotes
(1) In the present study, in line with other studies (e.g., Titone & Connine, 1994), idioms were
considered as predictable before offset if at least 70% of the idiom fragment completions were
idiomatic, and unpredictable if less than 20% of the completions were idiomatic (see Table 1 for the
actual mean cloze values).
(2) The authors found no evidence of semantic activation of the literal word meanings when
unambiguous idioms were presented in biasing contexts suggesting that top-down processes driven
by context and expectancies were sufficient to switch off word-level semantic processing.
Processing ambiguous idioms
28
Figure 1
Mean reading times (in ms) in the Disambiguating Region of Literal stories and Idiomatic stories
for Predictable and Unpredictable idioms.
1400
1450
1500
1550
1600
1650
1700
Literal Story Idiomatic Story
Predictable idioms Unpredictable idioms
Processing ambiguous idioms
29
Table 1
Mean ratings of Familiarity, Predictability, Meaning dominance, Age of Acquisition, Semantic
transparency, Story plausibility for Predictable and Unpredictable idioms, Cumulative word
frequency (log-transformed) in the Disambiguating region and statistical tests. Standard deviations
are reported in parentheses.
Predictable
idioms
Unpredictable
idioms
Familiarity*
5.86 (.5)
5.54 (.5)
t
(30) = 1.79, p = .083
Idiom Predictability
79.9%
(12.0%)
13.9%
(12.5%)
t (30) = 15.2, p = .000
Meaning dominance**
3.98 (.5)
3.9 (.4)
t
(30) = .496, p = .624
Age of Acquisition*
5.33 (.55)
5.25 (.47)
t
(30) = .434, p = .667
Semantic transparency**
2.88 (.6)
2.84 (.7)
t
(30) = .163, p = .871
Story plausibility*
4.54 (.84)
4.8 (1.06)
F
s
< 1
Log transf. cumulative
word frequency in the
Disambiguating region
7.94 (.43)
7.84 (.48)
t (30) = .606, p = .549
* On 7-point scale; **on 5-point scale
Processing ambiguous idioms
30
Table 2
Italian examples of Literal and Idiomatic Stories for Predictable and Unpredictable Idioms with
word-by-word English translations in parentheses. Slashes indicate the boundaries of the four
regions.
Predictable idioms
Rimboccarsi le maniche (to roll up one’s sleeves)
Literal story
La mamma ha visto che / Ada si era rimboccata le maniche / per potersi infilare i guanti di gomma /
e fare le pulizie. (Mom saw that / Ada had rolled up her sleeves / to shove the rubber gloves / and
do the cleaning.)
Idiomatic story
La mamma ha visto che / Ada si era rimboccata le maniche / per poter terminare la tesi di laurea /
entro tre settimane. (Mom saw that / Ada had rolled up her sleeves / to finish the dissertation /
within three weeks.)
Tenere sotto una campana di vetro (to keep under a glass bell)
Literal story
La signora sapeva che / Marta teneva sotto una campana di vetro / un antico vaso etrusco per
proteggerlo / dai possibili urti. (The lady knew that / Marta kept under a glass bell/ an ancient
Etruscan vase to protect it / from physical impact.)
Idiomatic story
La signora sapeva che / Marta teneva sotto una campana di vetro / il figlio per proteggerlo dalle
difficoltà / della vita quotidiana. (The lady knew that / Marta kept under a glass bell/ the son to
protect him from the troubles / of everyday life.)
Mettere i bastoni tra le ruote (to put the clubs in the wheels; to put a spoke in the wheels)
Literal story
Processing ambiguous idioms
31
La moglie si accorse che / Carlo aveva messo i bastoni tra le ruote / del carro per evitare che si
muovesse / giù per la discesa. (The wife realized that /Carlo had put a spoke in the wheels/ of the
chart to prevent it from moving/ down from the descent.)
Idiomatic story
La moglie si accorse che / Carlo aveva messo i bastoni tra le ruote / al collega per impedire che
fosse / promosso dal titolare. (The wife realized that /Carlo had put a spoke in the wheels/ of the
colleague to prevent him from being/ promoted by the boss.)
Unpredictable idioms
Giocare col fuoco (to play with fire)
Marisa raccontò che / il ragazzo aveva giocato col fuoco / quando aveva trovato dei fiammiferi per
terra / nel cortile della scuola. (Marisa said that / the boy had played with fire / when he found some
matches on the ground / in the schoolyard.)
Marisa raccontò che / il ragazzo aveva giocato col fuoco / quando aveva guidato in modo
spericolato / l’automobile del padre. (Marisa said that / the boy had played with fire / when he drove
recklessly / the father's car.)
Fare i salti mortali (to turn the somersaults; to make the impossible)
Literal story
La mamma ha detto che / Lisa aveva fatto i salti mortali / quando aveva ballato il rock acrobatico /
nello spettacolo della scuola (Mom said that / Lisa had turned the somersaults / when she danced
the acrobatic rock / in the show of the school.)
Idiomatic Story
La mamma ha detto che / Lisa aveva fatto i salti mortali / quando aveva dato quattro esami in un
mese / per rimanere in corso (Mom said that / Lisa had turned the somersaults / when she had taken
four exams in a month / to remain in due course.)
Tagliare la corda (To cut the rope, to snake off)
Processing ambiguous idioms
32
Literal story
La signora disse che / Sergio aveva tagliato la corda/ per poter aprire il delicato pacchetto/
proveniente dalla sua città (The lady said that/ Sergio had cut the rope / to open the delicate
package/ coming from his hometown.)
Idiomatic story
La signora disse che / Sergio aveva tagliato la corda/ per sfuggire alla volante della polizia /
arrivata dopo la rapina (The lady said that/ Sergio had cut the rope / to escape the police car/ arrived
after the robbery.)
Processing ambiguous idioms
33
Table 3
Mean reading times of Predictable and Unpredictable idioms in the Idiom Region, the
Disambiguating Region and the Final Region in Idiomatic and Literal Stories. Standard deviations
are reported in parentheses.
Predictable idioms
Unpredictable idioms
Literal Story
Idiomatic Story
Literal Story
Idiomatic Story
Idiom Region
1535 (239)
1591 (256)
Disambiguating
Region
1647 (264)
1535 (214)
1497 (221)
1577 (283)
Final Region
1073 (222)
1031 (198)
1069 (195)
1056 (191)
Processing ambiguous idioms
34
Appendix
Full list of the Italian experimental stories for predictable and unpredictable idioms, with word-by-
word English translations. The Literal story associated to each idiom is reported first followed by
the Idiomatic story. Slashes indicate the boundaries of each of the four regions.
Predictable idioms
1.Tirarsi la zappa sui piedi (to pull the hoe on one’s own feet; to cut one’s own throat)
I vicini si accorsero che / Franco si era tirato la zappa sui piedi / quando aveva lavorato la terra
nell’orto / davanti alla casa. (The neighbors noticed that / Franco had pulled the how on his feet /
when he worked the land in the garden / front of the house.)
I vicini si accorsero che / Franco si era tirato la zappa sui piedi / quando aveva offeso il vecchio
superiore / davanti ai colleghi. (The neighbors noticed that / Franco had pulled the how on his feet /
when he had offended the old boss/ in front of the colleagues.)
2.Non sapere che pesci prendere (to not know which fishes to buy; To do not know how to get on)
L’amica vide che / Marta non sapeva che pesci prendere / in pescheria per preparare la grigliata /
insieme ai colleghi. (The friend saw that / Marta did not know which fishes to buy / to prepare the
barbecue / with her coworkers.)
L’amica vide che / Marta non sapeva che pesci prendere / durante il compito per risolvere il
problema / di fisica teorica. (The friend saw that / Marta did not know which fishes to buy / during
the exam to solve the problem / of theoretical physics.)
3.Sfregarsi le mani (To rub one’s own hands, to be happy and satisfied)
Gli amici hanno detto che / Guglielmo si sfregava le mani / quando cercava di riscaldarsele nel
freddo / pomeriggio di fine novembre. (The friends said that / Guglielmo rubbed his hands / when
he tried to warm them up in the cold / afternoon in late November.)
Gli amici hanno detto che / Guglielmo si sfregava le mani / quando aveva vinto il premio
miliardario / alla lotteria di Natale. ( The friends said that / Guglielmo rubbed his hands / when he
won the billionaire prize / of the Christmas lottery.)
4.Essere tutti sulla stessa barca (to be all in the same boat; to share a destiny)
Il vecchio ha raccontato che / gli uomini erano tutti sulla stessa barca / quando furono trasportati
dagli scafisti / sulle coste pugliesi. (The old man said that / the men were all in the same boat / when
they were transported by the smugglers / on the Apulian coast).
Il vecchio ha raccontato che / gli uomini erano tutti sulla stessa barca / quando furono licenziati
dalla fabbrica / nel periodo di crisi. (The old man said that / the men were all in the same boat /
when they were fired from the factory / in the period of crisis.)
5.Ficcare il naso (to puke one’s nose; to snoop)
La nonna ha visto che / Luisa aveva ficcato il naso / nei fiori della vicina per annusare / il profumo
del gelsomino. (The grandmother saw that / Luisa had puked her nose / in the flowers of the
neighbor to smell / the scent of jasmine.)
La nonna ha visto che / Luisa aveva ficcato il naso / negli affari della vicina per sapere / della
vendita della casa. (The grandmother saw that / Luisa had puked her nose / in the affairs of the
neighbor to know / the sale of the house.)
6.Fasciarsi la testa (to bandage one’s own head; to anticipate a potentially unfortunate event)
La mamma ha detto che / Ugo si era fasciato la testa / quando si era ferito urtando lo spigolo /
dell’armadietto di metallo. (Mom said that / Ugo had bandaged his head / when he was hurt
bumping the edge / of the metal cabinet.)
La mamma ha detto che / Ugo si era fasciato la testa / prima di sapere il risultato dell’esame / di
storia della politica. (Mom said that / Ugo had bandaged his head / before knowing the result of the
exam / of history of politics.)
7.Mettere i bastoni tra le ruote (to put the clubs in the wheels; to put a spoke in the wheel)
Processing ambiguous idioms
35
La moglie si accorse che / Carlo aveva messo i bastoni tra le ruote / del carro per evitare che si
muovesse / giù per la discesa. (The wife realized that /Carlo had put a spoke in the wheels/ of the
chart to prevent it from moving/ down from the descent.)
La moglie si accorse che / Carlo aveva messo i bastoni tra le ruote / al collega per impedire che
fosse / promosso dal titolare. (The wife realized that /Carlo had put a spoke in the wheels/ of the
colleague to prevent him from being/ promoted by the boss.)
8.Rimanere sullo stomaco (to stay on the stomach; to cannot stand)
La suocera ha detto che / a Cesare era rimasta sullo stomaco / l’anatra all’arancia quando aveva
cenato / nel ristorante francese. (The mother-in-law said that / to Caesar remained on the stomach /
the orange duck when he had dinner / in the French restaurant.)
La suocera ha detto che / a Cesare era rimasta sullo stomaco / la brutta discussione quando aveva
litigato / col vicino arrogante . (The mother-in-law said that / to Caesar remained on the stomach /
the ugly discussion when he argued / with the arrogant neighbor.)
9.Cavarsi il dente (to extricate a tooth; to tell something unpleasant)
Lo zio ha detto che / Renata si era cavata il dente / quando era andata dal rinomato dentista / nella
clinica private. (The uncle said that / Renata had extricated the tooth / when she went to the
renowned dentist / in the private clinic.)
Lo zio ha detto che / Renata si era cavata il dente / quando aveva dato il difficile esame / di diritto
pubblico. (The uncle said that / Renata had extricated the tooth / when she had taken the difficult
examination / of public law.)
10.Battere il ferro finché è caldo (to strike while the iron is hot; to profit)
Gli amici videro che / Luigi doveva battere il ferro finché era caldo / per costruire la cancellata
dell’edificio / costruito di recente. (The friends saw that / Luigi had to strike while the iron was hot /
to build the gate of the building / built recently.)
Gli amici videro che / Luigi doveva battere il ferro finché era caldo / per ottenere la promozione dal
direttore / della filiale di Forlì. (The friends saw that / Luigi had to strike while the iron was hot / to
to be promoted as director / of the branch of Forli.)
11.Essere sulla cresta dell’onda (to ride the crest of the wave)
Il ragazzo ha raccontato che / Valentina era sulla cresta dell’onda / quando andava in surf ed è
scivolata / facendosi male al ginocchio. (The boy told that / Valentina was riding the crest of the
wave / when she was surfing and had slipped / hurting her knee.)
Il ragazzo ha raccontato che / Valentina era sulla cresta dell’onda / quando aveva recitato
nell’ultimo film / del famoso regista francese. (The boy told that / Valentina was riding the crest of
the wave / when she had starred in the last movie / of the famous French director.)
12.Tenere sotto una campana di vetro (to keep under a glass bell.)
La signora sapeva che / Marta teneva sotto una campana di vetro / un antico vaso etrusco per
proteggerlo / dai possibili urti. (The lady knew that / Marta kept under a glass bell/ an ancient
Etruscan vase to protect it / from physical impact.)
La signora sapeva che / Marta teneva sotto una campana di vetro / il figlio per proteggerlo dalle
difficoltà / della vita quotidiana. (The lady knew that / Marta kept under a glass bell/ the son to
protect him from the troubles / of everyday life.)
13.Stringere i denti (To grit the teeth; to bite the bullet)
La Signora Rossi ha detto che / Lucia aveva stretto i denti / per provare la dentiera appena fatta /
nello studio di Bologna. (Ms. Rossi said / Lucia had gritted the teeth / to try the test dentures just
made / in the study of Bologna.)
La Signora Rossi ha detto che / Lucia aveva stretto i denti / per terminare il lavoro entro sera / e
andare in ferie. (Ms. Rossi said / Lucia had gritted the teeth / to finish the job by evening / and go
on vacation.)
14.Strapparsi i capelli (to rip one’s hair; to be desperate)
Processing ambiguous idioms
36
Le amiche hanno detto che / Luisa si era strappata i capelli / quando si erano impigliati tra i rovi /
dei cespugli nel bosco. (The friends said they / Luisa had ripped the hair / when they were caught
between the brambles / of the bushes in the woods.)
Le amiche hanno detto che / Luisa si era strappata i capelli / quando aveva perso l’anello di brillanti
/ di grandissimo valore. (The friends said they / Luisa had ripped the hair / when she had lost the
diamond ring / of great value.)
15.Avere l’acqua alla gola (to have the water at the throat; to be in a tight corner)
La zia ha raccontato che / Giovanni aveva l'acqua alla gola / quando era caduto nel piccolo laghetto
/ in mezzo al giardino. The aunt said that / John had the water at the throat / when he felt in the little
pond / in the middle of the garden.)
La zia ha raccontato che / Giovanni aveva l'acqua alla gola / quando doveva pagare il grosso debito
/ agli spietati usurai. (The aunt said that / John had the water at the throat / when he had to pay the
huge debt / to ruthless loan sharks.)
16. Rimboccarsi le maniche (to roll up one’s sleeves)
La mamma ha visto che / Ada si era rimboccata le maniche / per potersi infilare i guanti di gomma /
e fare le pulizie. (Mom saw that / Ada had rolled up her sleeves / to shove rubber gloves / and do
the cleaning.)
La mamma ha visto che / Ada si era rimboccata le maniche / per poter terminare la tesi di laurea /
entro tre settimane. (Mom saw that / Ada had rolled up her sleeves / to finish the dissertation /
within three weeks.)
Unpredictable idioms
1.Essere in alto mare (to be on the high seas)
La fidanzata temeva che / il ragazzo fosse ancora in alto mare / quando iniziò una fortissima
tempesta / con tuoni e fulmini. (The girlfriend was afraid that / the boy was still on the high seas /
when a very strong storm started/ with thunder and lightning.)
La fidanzata temeva che / il ragazzo fosse ancora in alto mare / quando giunse la scadenza
definitiva / per consegnare il lavoro. (The girlfriend was afraid that / the boy was still on the high
seas / when he reached the deadline / to deliver the work.)
2.Fare il primo passo (to take one’s first step)
La mamma ha raccontato che / Stefano aveva fatto il primo passo / per cominciare a camminare da
solo / quando aveva dieci mesi. (Mom said that / Stefano had taken the first step / to begin to walk
alone / when he was ten months.)
La mamma ha raccontato che / Stefano aveva fatto il primo passo / per fare amicizia con la nuova
compagna / arrivata dall’Algeria. (Mom said that / Stefano had taken the first step / to make friends
with his new partner / arrived from Algeria.)
3.Scendere in campo (to take the field)
L’amico ha raccontato che / Andrea era sceso in campo / per partecipare alla partita di calcio / del
campionato della scuola. (The friend said that / Andrea had taken the field / to attend the football
game / at the championship of the school.)
L’amico ha raccontato che / Andrea era sceso in campo / per difendere i diritti dei lavoratori /
durante le lotte sindacali. (The friend said that / Andrea had taken the field / to defend the rights of
workers / during the labor struggles.)
4.Essere a cavallo (to be on the horseback; to be satisfied)
Luigi ha raccontato che / il giovane era a cavallo / quando aveva visto le guardie forestali / nel
parco naturalistico. (Luigi said that / the young man was on the horseback / when he saw the
Rangers / in the natural park.)
Luigi ha raccontato che / il giovane era a cavallo / quando era riuscito a superare l’esame / di
economia finanziaria. (Louis said that / the young man was on the horseback / when he managed to
pass the exam / of financial economics.)
Processing ambiguous idioms
37
5.Fare i salti mortali (to turn the somersaults; to make the impossible)
La mamma ha detto che / Lisa aveva fatto i salti mortali / quando aveva ballato il rock acrobatico /
nello spettacolo della scuola (Mom said that / Lisa had turned the somersaults / when she danced
the acrobatic rock / in the show of the school.)
La mamma ha detto che / Lisa aveva fatto i salti mortali / quando aveva dato quattro esami in un
mese / per rimanere in corso. (Mom said that / Lisa had turned the somersaults / when she had taken
four exams in a month / to remain in due course.)
6.Tirare la cinghia (to pull the belt; to sacrifice oneself)
La moglie capì che / Giulio aveva tirato la cinghia / per sorreggere i pantaloni ormai larghi/ dopo
essere dimagrito. (The wife understood that / Giulio had pulled the belt / to hold the pants baggy
now / after being slimmed.)
La moglie capì che / Giulio aveva tirato la cinghia / per poter comprare una nuova automobile /
senza chiedere il mutuo. (The wife understood that / Giulio had pulled the belt / in order to buy a
new car / without asking the mortgage.)
7.Tagliare la corda (to cut the rope; to slope off)
La signora disse che / Sergio aveva tagliato la corda / per poter aprire il delicato pacchetto /
proveniente dalla sua città (The lady said that/ Sergio had cut the rope / to open the delicate
package/ coming from his hometown.)
La signora disse che / Sergio aveva tagliato la corda / per sfuggire alla volante della polizia /
arrivata dopo la rapina. (The lady said that/ Sergio had cut the rope / to escape the police car/
arrived after the robbery.)
8.Essere sulla strada giusta (to be on the right ; to be on the right track)
Il babbo sapeva che / Guglielmo era sulla strada giusta / per raggiungere la casa degli amici / vicino
al Monte Bianco. (The father knew / Guglielmo was on the right road / to reach the house of his
friends / near the Mont Blanc.)
Il babbo sapeva che / Guglielmo era sulla strada giusta / per intraprendere una brillante carriera /
nella ricerca spaziale. (The father knew / Guglielmo was on the right road / to pursue a brilliant
career / in space research.)
9.Trovare la pappa pronta (to find the pap ready; to find everything on a plate)
La nonna ha detto che / Renato aveva trovato la pappa pronta / quando era tornato a casa dall’asilo /
alle sei del pomeriggio. (The grandmother said / Renato had found the pap ready / when he came
home from kindergarten / at six in the afternoon.)
La nonna ha detto che / Renato aveva trovato la pappa pronta / quando il padre l’aveva subito
assunto / nell’azienda di famiglia. (The grandmother said / Renato had found the pap ready / when
the father had immediately taken him / in the family business.)
10.Avere le mani legate (to have the hands tied; to be bound hand and foot)
La signora ha raccontato che / il Signor Rossi aveva le mani legate / quando i ladri l’avevano
immobilizzato / nel retro del negozio. (The lady said that / Mr. Rossi had the hands tied / when the
thieves had immobilized him/ in the back of the store.)
La signora ha raccontato che / il Signor Rossi aveva le mani legate / quando non poteva dire quello
che pensava / al suo superiore. (The lady said that / Mr. Rossi had the hands tied / when he could
not say what he thought / to his boss.)
11.Dare una strigliata (to give a dressing; to give a telling-off)
Il marito ha saputo che / Maria aveva dato una strigliata / al suo cavallo quando si era sporcato /
dopo la lunga cavalcata. (The husband knew / Maria had given a dressing down / to her horse when
it was soiled / after the long ride.)
Il marito ha saputo che / Maria aveva dato una strigliata / a suo figlio quando era stato bocciato /
all’esame di inglese. (The husband knew / Maria had given a dressing down / to his son when he
was rejected / at the exam of English.)
12.Giocare col fuoco (to play with fire)
Processing ambiguous idioms
38
Marisa raccontò che / il ragazzo aveva giocato col fuoco / quando aveva trovato dei fiammiferi per
terra / nel cortile della scuola. (Marisa said that / the boy had played with fire / when he found some
matches on the ground / in the schoolyard.)
Marisa raccontò che / il ragazzo aveva giocato col fuoco / quando aveva guidato in modo
spericolato / l’automobile del padre. (Marisa said that / the boy had played with fire / when he drove
recklessly / the father's car.)
13.Sentire puzza di bruciato (to smell a stink of burning; to smell a rat)
La domestica ha raccontato che/ Luisa aveva sentito puzza di bruciato / quando aveva strinato
l’arrosto di vitello / preparato per la cena. (The maid told that / Luisa had smelled a stink of burning
/ when he had burned the roast veal / prepared for dinner.)
La domestica ha raccontato che/ Luisa /aveva sentito puzza di bruciato / quando aveva visto delle
macchie di rossetto / sulla camicia del marito. (The maid told that / Luisa had smelled a stink of
burning / when she had seen the lipstick stains / on her husband's shirt.)
14.Buttarsi dietro le spalle (to throw over one own’s shoulders; to intentionally forget)
La zia ha visto che / Serena si era buttata dietro le spalle / i lunghi capelli castani per legarli / con un
nastro arancione. (Aunt saw that / Serena had thrown over her shoulders / the long brown hair to tie
them / with an orange ribbon.)
La zia ha visto che / Serena si era buttata dietro le spalle / il matrimonio fallito per ricominciare /
una vita più serena. (Aunt saw that / Serena had thrown over her shoulders / the marriage failed to
start / a more serene life.)
15.Giocare in casa (to play at home; to be advantaged)
La mamma disse che / Luigi aveva giocato in casa/ quando lei era andata al supermercato / a fare la
spesa. (Mom said that / Luigi had played at home / when she went to the supermarket / for grocery
shopping.)
La mamma disse che / Luigi aveva giocato in casa/ quando aveva vinto il concorso interno /nella
sua università. (Mom said that / Luigi had played at home / when he won the internal competition /
at his hometown university.)
16.Prendere di petto (to take for chest; to face up to something)
Il ragazzo ha visto che / Riccardo aveva preso di petto / la palla quando il giocatore avversario /
l’aveva lanciata in campo. (The boy saw that / Riccardo had taken for chest / the ball when the
opposing player / threw it in the field.)
Il ragazzo ha visto che / Riccardo aveva preso di petto / la discussione quando la madre arrabbiata /
l’aveva sgridato duramente. (The boy saw that / Riccardo had taken for chest / the discussion when
the angry mother / had scolded him severely.)
