How fixed are phrases? A linguistic model of fixedness and its
psycholinguistic application in primary school children.
Marianna Mini
Ph.D. (Educ),
University of Patras
Kleopatra Diakogiorgi Aggeliki Fotopoulou
Dept of Education,
Faculty of Humanities
& Social Sciences,
University of Patras,
Greece
Aggeliki Fotopoulou
Institute for Language
and Speech
Processing (ILSP),
Athens, Greece
1 Introduction
‘Phrases’, ‘idioms’ and ‘idiomatic expressions’ are some of the terms used
to refer to one of the commonest forms of figurative language. Yet, despite the
fact that for more than four decades idiomatic expressions have been under close
scrutiny, approached from various angles (syntax, semantics, pragmatics), no one
can strongly argue that the issue has been resolved. While this can be accounted
for in many ways, perhaps the major reason is that the linguistic phenomena
referred to in the above terms are not completely defined. A large number of fixed
expressions are characterized by the fact that their figurative meaning is not a
direct function of the meaning of their parts (Jackendoff, 1997). For example, the
composition of the individual meanings of the words kick, the, bucket do not
produce the meaning ‘die suddenly’. Consequently for the speaker, access to the
meaning of these phrases is not possible through the analysis of the meaning of
their component words.
Nevertheless, this is how past idiom research used to treat idiomatic
expressions (Fraser, 1970, Chomsky, 1980, Gross, 1982, 1988 and van der
Linden, 1992). According to more recent views, called compositional views
(Gibbs & Nayak, 1989; Gibbs, Nayak, Bolton, & Keppel, 1989) many idioms are,
at least in part, analyzable or decomposable word configurations. Nunberg (1978)
was the first to introduce the notion of idioms as combing expressions.
Compositionality refers to the fact that the constituents of some idioms “carry
identifiable parts of the idiomatic meaning” (Nunberg, Sag, & Wasow, 1994,
p.496). Based on this notion, in addition to the distinction between semantically
decomposable and non-decomposable idioms, Nunberg proposed an idiom
classification that emphasized the interaction between an idiom’s literal and
figurative meanings. According to this classification scheme, idioms vary in their
semantic decomposability. He distinguishes two types of decomposable idioms:
the normally decomposable idioms and the abnormally decomposable idioms.
Normally decomposable idioms’ constituents bear a direct relation to the
figurative meaning. In pop the question, for example, there is a clear
correspondence between pop and question and the relevant parts of the figurative
meaning ‘propose marriage’. Abnormally decomposable idioms’ constituents
bear a metaphorical relation to the figurative meaning. For example, the word
maker in meet your maker metaphorically refers to a deity.
In psycholinguistics, compositionality was introduced by Gibbs and colleagues
(Gibbs & Nayak, 1989; Gibbs, Nayak, Bolton, & Keppel, 1989). Based on the
notion of compositionality proposed by Nunberg, Gibbs and colleagues developed
the so-called idiom decomposition hypothesis. This hypothesis asserts that
semantically decomposable idioms may be analyzed compositionally. Each
component is retrieved from the mental lexicon and combined with the other
components of the string according to their syntactic relations. In contrast, the
meaning of non decomposable idioms is retrieved directly from the lexicon. Thus,
decomposable and non decomposable idioms are represented differently and
processed in different ways. With respect to comprehension, the main prediction
of compositional models is that an idiom will be easier to comprehend if its words
are related, in any way, to its figurative meaning. In spill the beans - the most
commonly cited example of a decomposable idiom in the literature- it is assumed
(Nayak, 1989; Gibbs, Nayak, Bolton, & Keppel, 1989; Gibbs, Nayak, & Cutting,
1989; Hablin & Gibbs, 1999), that there is an isomorphism between the individual
constituents’ meaning and the idiomatic meaning of the expression as a whole.
Every constituent contributes – literally or metaphorically- to the idiomatic
meaning of the expression, with beans corresponding to the secrets and spill to the
action of divulging the secrets. Decomposable idioms have been combined with
syntactic as well as processing properties: greater flexibility, accuracy and rapidity
of access. Yet, the empirical evidence in support of these claims is not always very
consistent (see Tabossi, Fanari & Wolf, 2008).
Past research treated idioms as word configurations where there is a slight
relationship between literal and figurative meaning. Therefore, knowing the literal
meaning of an idiom does not aid the comprehender to find its figurative meaning.
(Orthony, Schallert, Reynolds & Antos, 1978). This led to the hypothesis that
language users (children also) learn idioms as giant lexical units and not by
analysing their constituents (Ackerman, 1982). Nippold and Rudzinski (1993) are
totally in disaccord with this (1993). According to her hypothesis, idiom
comprehension is largely facilitated when language users are capable of analysing
them semantically. Transparent idioms provide an excellent opportunity for such
an analysis given that the literal meaning of these idioms may help to cue their
figurative meaning (Caillies & Le Sourn-Bissaoui, 2006; Nippold & Duthie,
2003). Adults and adolescents are sensitive to differences in transparency and can
reliably rate the contribution of key words to the meanings of idiomatic
expressions (Titone & Connine, 1999). When children acquire a similar sensitivity
is not clear. Findings about the age at which semantic analysis can be used
successfully to process an idiom’s meaning are conflicting. According to several
studies, this is an early developing skill. For example, Gibbs (1987, 1991) found
that children as young as 5 year-olds are better at explaining transparent idioms
than opaque ones. Other findings seem to support a late development of this skill.
For example, Nippold and Rudzinski (1993) failed to find a positive correlation
between transparency and performance on an idiom explanation task for 11-yearolds but they found a positive one for 14 and 17- year-olds. Similar findings were
obtained by Nippold and Taylor (2002). According to Levorato and Cacciari
(1999), semantic analysis can by used by children as these acquire a processing
skill that, nevertheless, develops with age. Thus, it seems that the ability to analyse
the internal semantics of an idiom needs some time to develop.
The present study aims at investigating the extent to which idiomatic phrases
are fixed, through a thorough linguistic analysis of 430 phrases with fixed subject
in Greek. A linguistic model of fixedness is developed, the empirical adequacy of
which is assessed in elementary school aged children. One of the major aims of
this paper is to test 7.5 and 11.5 year-old Greek children’s sensitivity to semantic
parameters of idiom processing in their language. A series of experiments was
conducted, one of which will be presented in this paper. Linguistic material used
in these experiments was selected from a set of 450 phrases having a fixed relation
between verb and subject. In order to select the most appropriate phrases the
following procedure was followed. First search and compare the relative meanings
as these appeared in various dictionaries, and then search and find examples of
their use in linguistic contexts. The following sources were used for this purpose:
The Dictionary of Modern Greek (G. Babiniotis, 2002); The Dictionary of
Common Greek (M. Triantafyllidis Foundation, 1998), The Hellenic National
Corpus (ILSP, http://hnc.ilps.gr/). After having been collected, phrases were
classified according to syntactic and semantic criteria and indexed in a lexicon
accompanied by their syntactic and semantic properties. These criteria were the
same criteria as those used in the past in the framework of Lexicon-Grammar for
the classification of idioms in French by M. Gross (1975) and in Greek by
Fotopoulou (1993). The limits of this paper do not allow us to present in detail the
syntactic and the semantic properties of the idioms studied. We will just mention
three basic conclusions drawn by Gross.
1) Idiom fixedness can be limited to only certain constituents of the sentence.
The French sentence Max ne porte pas Luc dans son coeur (Max does not like
Luc) demonstrates a combination of fixed and non fixed elements. Fixedness of
this expression relies on the relation of the verb (ne pas porter) (not carry) and
the prepositional phrase (dans son coeur) (in his heart). The first component Luc
is not fixed.
2) There are small groups of fixed constructions with similar meanings,
allowing some degree of element variation within the idiomatic expression, e.g. in
French Max (a raté+ a loupé + a manqué) le coche (Max missed his chance) or
Max a perdu (la boule + la boussole + le Nord + les pedales…) (Max flipped
out). This partial flexibility, despite the fact that the meaning of these sentences is
not derived from the meaning of individual elements, made G. Gross (1996) speak
about degree of fixedness (degré de figement), arguing that “variations are more
frequent than actual fixedness”.
3) fixed sentences are set upon a continuum, starting from free structured
combinations and ending with fixed expressions specified as prototypical, i.e.
semantically opaque and structurally fixed. For example: Max a (manqué + loupe
+ raté) une chance unique
+ un beau coup
+ une bonne occasion
+ une affaire
+ le coche
In other words, fixed and free expressions are not so different. The detailed
description of the syntactic properties of phrases with a fixed subject led us to the
same conclusion. Nevertheless, concerning the lexical distribution of the phrases,
we decided to adopt a much more qualitative rather than a quantitative
perspective, going beyond the two basic criteria used by Gross. These are the
semantic criterion according to which the meaning of an idiom is not derived
from the meaning of its parts, and the lexical–structural criterion according to
which one or more elements of the clause is lexically invariable in relation to the
verb. Certainly, the criterion of restricted distribution is a crucial one for
excluding all expressions that contain verbs with a large distribution coming from
various semantic fields. However, the categorization of the phrases as to the
degree of their fixedness needs more subtle criteria than the purely quantitative
and finally insufficient criterion of restricted distribution. Thus, in each
expression we examined to what extent its meaning derives exclusively from the
lexical relationship between its verb and its noun or, on the contrary, the extent to
which the same verb with another noun can retain the meaning it has in this
expression. For example, in the expression δεν πέφτει καρφίτσα [literally ‘a pin
doesn’t drop’], which means ‘there are a lot of people crowded’, neither the verb
nor the noun can be found in another lexical configuration with the same or a
similar meaning to which they have in the expression δεν πέφτει καρφίτσα [‘a pin
doesn’t drop’].
Using the aforementioned test as a starting point, a spectrum of fixedness was
then created. This spectrum includes two types of phrases: The so-called “typical
phrases” and the “non typical phrases”. Non typical phrases include “quasi
phrases” and “conventionalized phrases”. In order to be characterized as typical, a
phrase has to meet the exclusive co-occurrence criterion – i.e. the verb constituent
of the phrase cannot maintain its meaning if it co-occurs with a noun constituent
outside the semantic field to which the noun of the recurrent phrase belongs. In
typical phrases, the semantic-lexical articulation between their constituents is so
strong that semantic autonomy is restricted or even completely neutralized. A non
typical phrase is one that meets the non exclusive co-occurrence criterion. In
quasi phrases, the verb retains (a) its literal meaning, or (b) the meaning it has at
the tangible/concrete level, or alternatively (c) is used with a non-literal meaning.
For example, in the phrase τον φοβήθηκε το µάτι µου [literally ‘my eye feared
him’] meaning ‘I was scared stiff by him or his actions’ the verb retains its literal
meaning (see, Η Έλλη τον φοβήθηκε [Elli feared him]). In the phrase πάγωσε το
αίµα της [her blood froze] meaning ‘she was shocked’ the verb has a non-literal
meaning (see ο Άρης πάγωσε [Aris was frozen] meening ‘Aris was shocked’). In
the phrase ράγισε η καρδιά µου [my heart cracked], the verb retains the
fundamental meaning of ‘cracking without being cut into pieces’ it has in the
sentence ράγισε το ποτήρι [the glass cracked] by a semantic extension on a non
tangible/abstract level.
Conventionalized phrases are those where the verb maintains its fundamental
meaning whereas the noun has a parallel to the basic meaning, a conventionalised
meaning. In the phrase τον τρώει το άγχος [literally ‘the stress eats him’ meaning
stress wears him out] the verb eat retains the core meaning of wear. What
differentiates this phrase from ‘ο σκόρος τρώει το µάλλινο ύφασµα’ [literally ‘the
moth eats the woollen material’ meaning moths eat wool] is that what the phrase
τον τρώει το άγχος describes, is realized on an abstract level, whereas its
consequences are perceived on a tangible level. In other words, one can recognise
somebody who is anxious, but one can hardly see him “eaten alive” by stress. An
important difference between conventionalized phrases and quasi phrases is that
in the first, noun constituents are not parts of the body, and therefore they do not
refer to human behaviour through a metonymic relationship part/organ –
whole/carrier.
In conclusion, typical phrases, quasi phrases and conventionalized phrases are
three discreet categories of idioms that form the spectrum of fixedness. However,
a crucial question for those interested in psycholinguistics is to what extent this
graded-fixedness model has a psychological reality or, on the contrary, to what
extent it is a model with a merely theoretical, lexicographical application.
Accordingly, the main goal of the experiment presented below is to explore
children’s sensitivity to those -sometimes subtle- semantic parameters, by testing
the effect of the phrase category on the access to the idiomatic meaning. This
research aims at evaluating the psychological validity of the graded fixedness
model we proposed based on the notion of semantic analyzability. Semantic
analyzability is the only relevant factor in this experiment. Therefore, we tried to
control-out two of the factors most likely to interfere, namely familiarity and
context. Idioms do differ widely in their frequency of occurrence from one
language to another. There is also strong empirical evidence in support of the
view that familiarity with idioms is has a role to play in idiom comprehension,
and especially so for younger children (Ortony, Turner, and Larson – Shapiro,
1985; Popiel & McRae, 1988; Schweigert, 1986).
In order to establish the non-familiarity of the linguistic material to the
participants, an additional preliminary research was carried out, in which the
familiarity of 33 phrases was tested. Twenty-two (22) fifth graders, attending
elementary schools in Athens participated in this additional preliminary research.
Bearing in mind that familiarity with any expression is hard to be ruled-out in the
first place, the basic idea was to control it assuming that there is an agedependency in the understanding of any phrase. It is only reasonable to assume
that if phrases were found to be unfamiliar for the older students, these would be
even more so for the younger ones. Therefore, we chose for the purpose of the
present study, to conduct this additional preliminary test using fifth-graders, these
being older than our main student-sample of second-graders and fourth-graders.
In the first phase of this familiarity test, children were asked to complete a
questionnaire where they were questioned about each phrase of the material,
whether they had heard it before (henceforth called frequency test). In the second
phase of the preliminary research, children were presented with a booklet
containing the same 33 phrases. A multiple–choice test was given to them
(henceforth comprehension test). They had to choose, from three interpretations
(idiomatic interpretation, literal interpretation, other interpretation) the one they
judged the most appropriate. The interpretations we have called “other” are
plausible answers but not connected to the phrases either lexically or
conceptually. Statistical analysis showed that, except for the case of four
particular phrases, there was no significant correlation between the two tests.
Thus, the material selected for the main research consisted of the phrases that
gathered the lowest scores in the comprehension test.
As far as context is concerned, there is growing evidence in support of its
crucial effect on idiom comprehension. An early sensitivity to context in language
learning and comprehension is well documented. Thus, it is not surprising to find
that children aged 7 and above understand idiomatic expressions better when
these are embedded in a supportive context. This is in line with previous
investigations into idiom comprehension (e.g. Levorato & Cacciari, 1995, 1999).
According to the working hypothesis tested by the present study, the greater
the semantic autonomy of the phrase’s constituents, the better the comprehension
is. In particular, we expected that typical phrases whose constituents have
restricted semantic autonomy would be more difficult for children than nontypical phrases.
Concerning non-typical phrases, we expected that quasi phrases would be
more difficult for children than conventionalized phrases. Conventionalized
phrases were expected to be more accessible than quasi phrases, as their nominal
constituents have tangible referents and that their content can be actualized. In
theory, the parts of the body to which the nominal constituents of the quasi
phrases refer are also tangible. However, the nominal constituents of the
conventionalized phrases refer to concrete nouns which seem to be more
imageable for young children (see, Cacciari & Levorato, 1998).
2. Method
2.1 Participants
A total of two-hundred (200) Greek elementary school children participated in
the experiment. One hundred of these were second graders (aged from 7 years and
1 month to 7 years and 11 months, having mean age of 7 years and 6 months - 56
boys and 44 girls). Another one hundred of these were forth graders (aged from 9
years to 10 years, mean age of 9 years and 5 months - 41 boys and 59 girls). All
children attended elementary schools in Athens. None of these children was
dyslexic, nor had cognitive impairments or severe learning difficulties.
2.2 Material
A total of 24 phrases (8 typical, 8 quasi and 8 conventionalized phrases) were
selected. The phrases were presented out of context, in random order.
2.3 Procedure
Children were tested individually. At the beginning, the experimenter gave
them some basic instructions concerning the task, as well as three examples in
order to familiarize them with the task. The children’s task was to read each
phrase carefully and to choose from three interpretations proposed, (idiomatic
interpretation, literal interpretation, other interpretation) the one they judged
correct. The experiment took place during the second semester of the school year
so as the second graders would already have gained enough experience with
reading.
3. Results
The mean percentages of correct and wrong answers chosen by children from
the two Age Groups are presented in Table 1. According to the x2 test, there was
a significant dependence of Age Group and Type of Answer (χ2=51.573, p=0.000,
df=1): the younger the children were, the greater the number of wrong answers
(literal and other). In particular, percentages of correct answers were lower than
percentages of wrong answers for younger children (32.6 vs. 67.4). The reverse
pattern was observed for older children (56.2 % vs. 43.8% for the correct and
wrong answers, respectively). Detailed data analysis demonstrated that
percentages of correct (idiomatic) answers were, for 20 out of the 24 phrases,
significantly lower than those for wrong answers (literal and other) and, in any
case, extremely low (25.5%). In only two phrases, percentages of correct answers
were lower than percentages of wrong answers: Την Κυριακή ανοίγει η αυλαία για
το πρωτάθληµα, meaning “The curtain goes up for the champiomship” , τον
φοβήθηκε το µάτι µου meaning “I was scared stiff by him or his actions”. Possible
explanations of the reasons why these two phrases were accessed successfully
could be that the first one refers to a favourite subject for children, football,
whereas the second is a phrase where the verb retains its literal meaning.
These results suggest that children of this age level find it difficult to
understand the meaning of the phrases belonging to the spectrum of fixedness.
As far as forth graders’ performances are concerned, percentages of correct
answers were significantly higher than percentages of wrong answers for 9 out of
24 phrases. For three phrases, percentages of wrong answers were higher than
percentages of correct answers whereas for the remaining 12 phrases, no
significant difference was found. These results suggest that older children are
indeed better at understanding the meaning of the phrases proposed. One
interesting finding worth noticing is that the 80% of children’s choices at both age
levels are either idiomatic or literal. Children’s choices which have been
collectively grouped as “other” (corresponding to plausible answers but not
connected to the phrases either lexically or conceptually) are measured to be
marginal in size. Ιn other words, interpretations which seem to bear no obvious
relationship to the lexical data of the phrase are less acceptable by children. Taken
together, these results suggest that children try to get access to the meaning of the
phrase either literally or non literally.
Age group
Type of answers
Correct answers
Wrong answers
Second graders
33
67
Forth graders
56
44
Table 1. Percentages of correct and wrong answers for both age groups
More interesting for the aim of the present study was the significant effect of
the phrase category on the type of answers given by second and forth graders. A
2x3 (i.e. Age x Phrase category) analysis of Variance (ANOVA) was applied. The
results showed that, for both age groups, percentages of idiomatic answers
differed significantly depending on phrase category: for the second graders
(F=10.403, p=0.000) and for the forth graders F=11.612, p=0.000). Similarly, the
results showed that, for both age groups, percentages of literal answers differed
significantly depending on phrase category: for the second graders (F=10.403,
p=0.000) and for the forth graders F=11.612, p=0.000).
More specifically, for the second graders, significant differences were
observed for both idiomatic and literal answers, for the comparison between 1)
typical phrases and quasi phrases and 2) typical phrases and conventionalized
phrases whereas between quasi phrases and conventionalized phrases no
significant difference was found (see Table 2). According to the results, children
of this age group find it more difficult to understand the meaning of the phrases
the components of which do not have a semantic autonomy than when they do.
Note however, the absence of a significant difference between quasi and
conventionalized phrases. That suggests that children of this age group are not yet
sensitive to subtle semantic differentiations of the linguistic material used in this
experiment. For the forth graders, significant differences were observed, for both
idiomatic and literal answers for the comparison between 1) typical phrases and
conventionalized phrases and 2) quasi phrases and conventionalized whereas
between typical phrases and quasi phrases no significant difference was found
(see Table 2). These results suggest that children of this age group are sensitive to
the semantic distinction between typical and non - typical phrases even if the
difference between typical phrases and quasi does not reach significance.
Furthermore, the significant advantage of conventionalized phrases over quasi
phrases reflects a refined sensitivity of older children compared to the younger
ones.
Age group Phrase category Type of answers Idiomatic literal other Second graders Typical 26 56 18 Quasi 34 45 21 Conventionalized 38 41 21 Forth graders Typical 50 29 21 Quasi 54 28 18 Conventionalized 65 20 20 Table 2. Percentages of idiomatic, literal and other answers chosen according to
age group and phrase category
4. Discussion
Two objectives were set at the beginning of the present study. The first
objective was to contribute to the investigation of the most crucial question in the
idiom literature, which is the question of fixedness. For this purpose, a semantic lexical categorization of 430 Greek phrases was attempted, those having a fixed
relation between verb and subject. Two major categories were distinguished,
based on the strength of the semantic - lexical articulation between verb and
subject: typical phrases and non - typical phrases. Typical were considered to be
the phrases that meet the exclusive co-occurence criterion: the verb does not cooccur with other noun constituent, while at the same time maintaining the
meaning it has in the phrase (e.g., in δεν πέφτει καρφίτσα [literally a ‘pin doesn’t
drop’] meaning ‘there are a lot of people crowded’). Conversely, non typical
phrases are those the constituents of which have a certain semantic autonomy.
Non-typical phrases were further categorized in: (a) quasi phrases and (b)
conventionalized phases. Quasi phrases are those where the verb maintains either
its literal or non-literal meaning (e.g., τον φοβήθηκε το µάτι µου [literally ‘my eye
feared him’] meaning ‘I was scared stiff by him or his actions’) or alternatively
maintaining its meaning at a tangible/concrete level (e.g., ράγισε η καρδιά της,
[literally ‘her heart cracked’] meaning ‘she was heart broken’). Conventionalized
are those where the verb maintains its fundamental meaning whereas the noun has
a second meaning parallel to the basic one, a conventionalised meaning (e.g. Τον
τρώει το άγχος της καθηµερινότητας [literally ‘the stress of everyday life eats
him’] meaning ‘stress wears him out’).
The second objective of the present study was to assess the psychological
reality of the graded fixedness model proposed. To this end, we investigated the
extent to which Greek elementary school children– aged between 7.5 and 9.5 are
sensitive to the aforementioned subtle semantic distinctions of idiomatic phrases
of their language, as this is identified by linguistic analysis. Given that the
linguistic material consisted in phrases controlled as to the parameter of
‘familiarity’, we assumed that the greater the semantic autonomy of the
constituents of the phrase, the more likely it is for the children to have access to
its idiomatic meaning.
The results provided further evidence in support of this hypothesis, showing
that in general, younger children (7.5 year-olds) have serious difficulties in
understanding the meaning of the phrases belonging to the spectrum of fixedness,
although they did perform better when treating non typical phrases than typical
ones. As a corollary to the above, older children (9.5 year-olds) also seem to
understand better non typical phrases than typical ones, although children of this
age group do exhibit significantly better performance, in general, than the
younger ones. Nevertheless, a noteworthy aspect of children’s phrase - processing
is that, with age, their strategies become not only more efficient but more
sophisticated as well. This can account for the significant advantage of
conventionalized phrases over quasi phrases being observed only at the age of 9.5
year-old group.
Observed overall, the results of the present study indicate that at the age of 7.5
year-olds semantic analysis can indeed be used by children to process phrases’
meaning, even though their performances can be quite poor. As children get older,
the use of this processing skill increases and indeed, it becomes more and more
refined. These results are consistent with a body of work (Gibbs, 1987, 1991;
Levorato and Cacciari, 1999; Nippold & Rudzinski, 1993; Nippold & Taylor,
2002) according to which children are able to analyse the internal semantics of the
phase in order to understand them. Naturally, further investigation is needed in
this area and particularly with regard to the most controversial question in this
research domain, namely the age of acquisition of this processing capacity.
References
Ackerman, B.P. 1982. On comprehending idioms: Do children get the picture? Journal of
Experimental Child Psychology. 33. 439-454.
Cacciari, C. & M.C. Levorato. 1998. The Effect of Semantic Analyzability of Idioms in
Metalinguistic Tasks. Metaphor and Symbol. 13 (3). 159-177.
Caillies S. & S. Le Sourn-Bissaoui. 2006. Idiom comprehension in French children: A
cock-and-bull story. European Journal of Developmental Psychology. 3 (2). 189206.
Chomsky, N. 1980. Rules and Representations. New York: Columbia University Press.
Fraser, B. 1970. Idioms within a Tranformational Grammar. Foundations of language. VI
(1). 22-42.
Fotopoulou, A. 1993. Une classification des phrases à compléments figés en grec
moderne - étude morphosyntaxique des phrases figées. Thèse de Doctorat. Paris:
Université Paris VIII - St. Denis.
Gibbs, W. R. 1987. Linguistic factors in children’s understanding of idioms. Journal of
Child Language. 14. 569-586.
Gibbs, W. R.1991. Semantic analysability in children’s understanding of idioms. Journal
of Speech and Hearing Research. 34. 613-620.
Gibbs, W. R., N. Nayak. 1989. Psycholinguistic studies on the syntactic behavior of
idioms. Cognitive Psychology. 21. 100-138.
Gibbs W. R.., N. Nayak, J. Bolton & M. Keppel. 1989a. Speakers' assumptions about the
lexical flexibility of idioms. Memory & Cognition. 16. 58-68.
Gibbs W. R., N.P. Nayak & C. Cutting. 1989b. How to Kick the Bucket and not
Decompose: Analyzability and Idiom Processing. Journal of Memory and
Language. 28. 576-593.
Gross, M. 1975. Méthodes en syntaxe. Paris: Hermann.
Gross, M. 1982. Une classification des phrases “figées” du français. In Revue Québecoise
de Linguistique. 11.2. 151-185.
Gross, M. 1988. Les limites de la phrase figée. In Langages. 90. 7-22. Paris: Larousse.
Hamblin, J., Gibbs, W.R.1999. Why You Can’t Kick the Bucket as You Slowly Die:
Verbs in Idiom Comprehension. Journal of Psycholinguistic Research. 28 (1).2539.
Ίδρυµα Μανόλη Τριανταφυλλίδη. 1998. Λεξικό της Κοινής Νεοελληνικής. Θεσσαλονίκη:
Ινστιτούτο Νεοελληνικών Σπουδών, Ίδρυµα Μανόλη Τριανταφυλλίδη.
Jackendoff, R. 1997. The Architecture of the Language Faculty. MIT Press.
Levorato, M. C., C. Cacciari. 1995. The effect of different tasks on the comprehension
and production of idioms in children. Journal of Experimental Child Psychology.
60. 261-283.
Levorato M.C. & C. Cacciari (1999). Idiom Comprehension in Children: Are the Effects
of Semantic Analysability and Context Separable?. European Journal of
Cognitive Psychology. 11 (1). 51-66.
Μπαµπινιώτης Γ. 2002. Λεξικό της Νέας Ελληνικής Γλώσσας, 2η έκδ. Αθήνα: Κέντρο
Λεξικολογίας.
Nippold, M. A., M. Rudzinski. 1993. Familiarity and transparency in idiom explanation:
A developmental study of children and adolescents. Journal of Speech and Hearing
Research. 36. 728-737.
Nippold, M. A.; C. L. Taylor. 2002. The judgments of Idiom Familiarity and
Transparency: A Comparison of Children and Adolescents. Journal of Speech,
Language and Hearing Research. 45.(2). 384-391.
Nippold M.A. & J.K. Duthie. 2003. Mental Imagery and Idiom Comprehension: A
Comparison of School-Age Children and Adults. Journal of Speech, Language,
and Hearing Research. 45. 788-799.
Nunberg, Geoffrey; Ivan Sag; Thomas Wasow. 1994. Idioms. Language. 70. 491-538.
Ortony A., D.L. Schallert, R.E. Reynolds & S.J. Antos. 1978. Interpreting Metaphors and
Idioms: Some Effects of Context in Comprehension. Journal of Verbal Learning
and Verbal Behaviour. 17. 465-477.
Ortony, T., and T. J. Turner and N. Larson-Shapiro. 1985. Cultural and Instructional
Influences on Figurative Language Comprehension by Inner City Children
Research in The Teaching of English.19 (1). 25-36.
Popiel, S,J., K. McRae. 1988. The figurative and literal senses of idioms, or all idioms are
not used equally. Journal of Psycholinguistic Research.17. 475-487.
Schweigert, W. 1986. The comprehension of familiar and less familiar idioms. Journal of
Psycholinguistic Research. 15. 33-45.
Tabossi P., R. Fanari, and K. Wolf. 2008. Processing Idiomatic Expressions: Effects of
Semantic Compositionality. Journal of Experimental Psychology: Learning,
Memory, and Cognition. 34 (2). 313-327.
Titone, D.A., C. M. Connine. 1999. On the compositional and noncompositional nature
of idiomatic expressions. Journal of Pragmatics. 31. 1655-1674.
The Hellenic National Corpus (ILSP), http://hnc.ilps.gr/).
Van der Linden E.J. 1992. Incremental processing and the hierarchical lexicon».
Computational Linguistics. 18. 219-238.