Traditional and Non-Traditional Approaches to Idioms

Toader M. Corina-Alexandra, Universitatea din Craiova, Facultatea de Litere

Traditional approaches to idioms

Structural and generative approaches will be considered as ‘traditional approaches’. According to

Graddol (1993), mainstream or traditional linguistics can be regarded as a manifestation of
structuralism. In this approach (which arguably inspired the generative model), language is viewed as an
autonomous mechanism (Graddol, 1993). This view is the common denominator between these
mainstream or traditional approaches. Traditional accounts of idioms focus on two problematic areas
which are mostly concerned with the generativist work of idiomaticity: how to represent idioms in the
speaker's lexicon, and how to account for their syntactic behaviour.

The structural approach

Structural linguistics views language as a systematic structure comprising ‘signifier’ and ‘signified’, i.e.
pairs of ‘thought’ and ‘sound’ in the mind of the speaker, as in Saussure’s approach. Along with other
cognitive linguists, Evans and Green (2006, p. 476) propose that both signified and signifiers are
“psychological entities”. They also mention that although there are differences between Saussure’s
approach and the cognitive linguistics approach, the cognitive approach nonetheless adopts the idea of
the Saussure an ‘symbol’. One of the broadest definitions of idioms from within this approach was put
forward by Hackett (1958). He espoused a structuralist-inspired model whereby any stretch of language
whose meaning cannot be deduced from the constituent parts is an idiom. His definition regards bound
morphemes such as the prefix pre- as the smallest unit of idioms. He claims that pre-is an idiom because
we cannot figure out its meaning from its structure. His definition fails to distinguish single lexical items
from grammatical structures that include multiple lexical items. For an empirical study on idioms,
making such a distinction is nonetheless useful, which probably helps explain why Hackett’s definition
has not been widely adopted.

In a structural model, language operates based on structural arrangements determined by abstract rules
(Holliday, 2006) which do not capture ‘idiosyncratic’ grammatical, lexical, or phraseological
phenomena (i.e. formulaic sequences). Chomsky (1972) was one of the first linguists to criticize
structural linguistic approaches for their failure to explain the complex and ambiguous sentence
structures of which formulaic sequences constitute a great part. His particular criticism was that
structural linguists do not adequately provide an efficient approach to account for language at the
phonological and morphological levels.
Both of these, he claims, have a finite number of units. However, he postulated that structural linguistics
failed to provide sufficient arguments to account for syntactic relations (in which the category of idioms
falls), as an infinite number of sentences can be generated, thus making it impossible to account for all
possible generated sentences. He proposed a set of rules that govern generating an infinite number of
sentences in the language. Chomsky’s dissatisfaction with the assumptions held by structural linguistics
to explain syntactical complexity was the impetus that led him to establish the foundations of generative
grammar.
The generative approach

Although the majority of early generative work in grammar focused on syntactic relations, the theory
was later applied to phonological and semantic components. Chomsky’s dissatisfaction with the
structuralist approach was over accounting for the syntactic relations where idioms fall. Idioms have
been a point of controversy in generative approaches in Standard Theory of transformational grammar
(Fraser, 1970; Katz and Postal 1963), and Generative Semantics (Newmeyer, 1972). It is fair to say,
then, that they pose a challenge to generative theories of syntax too. Because of space constraints, I will
consider only the transformational grammar approaches in relation to idioms.

Accounting for idioms was a major topic addressed by linguists working from a generative-inspired
approach. Katz and Postal (1963) presented a narrower definition than Hockett (1958). Their definition
integrates idioms into a transformational generative grammar model. Their approach assumes that the
whole meaning of an idiomatic expression or sentence that contains an idiomatic expression is different
from the meaning of the idiom’s constituent parts. In their view, then, polymorphemic words whose
compositional meanings are different from the meanings of their constituent parts form an idiom. A
word such as classroom, for instance, is regarded as an idiom because it refers to a ‘study room’ in a
school, which is different from the words class and room, which each have multiple meanings.
However, they differentiate classroom from words such as untrue, whose meaning is the same as the
meaning of un- + true. Katz and Postal’s definition of idioms also considers phrases whose meaning
cannot be retrieved from the individual words, as in kick the bucket. Put simply, any linguistic structure
whose meaning cannot be retrieved from its constituent elements qualifies as an idiom in such approach.

Katz and Postal (1963) distinguished ‘lexical idioms’ composed of polymorphemic words from ‘phrase
idioms’, whose structure falls into higher syntactic categories (phrases, clauses and sentences). Lexical
idioms belong to the lowest syntactic categories (noun, verb, adjective). They also point out that lexical
idioms are stored in a person’s lexicon and state that, unlike lexical idioms, phrase idioms are stored in
an ‘idioms list’, because an idiom has two possible meanings: literal or idiomatic. The literal meaning of
kick the bucket, for instance, is ‘to strike the bucket with a foot’, while the idiomatic meaning is ‘to die’.
Their criteria for phrase idioms are primarily concerned with productivity and transformability of the
expression. From their point of view, if an idiom allows transformation to the passive without losing its
figurative meaning – which e.g. kick the bucket does not – then it does not qualify as a phrase idiom,
because it is not productive.

Makkai’s (1972) investigation into idioms is one of the more influential ones, albeit from a formal
perspective.
Makkai (1972) situated idioms within a structural model that views polymorphemic words as idioms,
similar to Katz and Postal (1963). The only difference between the definition of Makkai and Katz and
Postal is that the latter includes words consisting of one bound and one free morpheme in their definition
of idioms, whereas Makkai (1972) only views polymorphemic words consisting of at least two free
morphemes (e.g. blackmail) as idioms.

Weinreich (1969) formulated a more applicable and even narrower definition. Only expressions such as
pull someone’s leg are considered idioms, according to his definition, whereas single words such as
telephone are not, thus overcoming the inapplicability of Hockett (1958) and Katz and Postal (1963). He
agrees with Katz and Postal that only expressions whose meaning cannot be retrieved from their
constituent parts are idioms. However, he limits the scope of idioms by using the term to refer only to
expressions that have both a literal and an idiomatic or figurative interpretation, as he believes that only
expressions causing ‘ambiguity’ as in pull someone’s leg qualify as idioms.

Weinreich (1969) also makes a strict distinction between idioms and ‘stable collocations’, such as part
and parcel, which are merely a co-occurrence of words because there is nothing idiomatic about them.
He views idiomaticity as causing ambiguity.
In his view, the co-occurrence of two words does not qualify an expression to be idiomatic, as it does not
involve a figurative or metaphorical usage – a characteristic of idiomaticity − which is at odds with
Firth’s (1957) notion of idioms and collocations. Firth treats idioms as irregular units from a semantic
perspective, but his focus is on their transformational deficiencies and ambiguity, which are important,
but are not the exclusive defining features. Fraser (1970) does not share Weinreich’s view. He sees any
‘constituent’ or a series of constituents whose semantic interpretation is not a compositional function of
the formatives as an idiom. In his model, idioms fall into a hierarchical system that chiefly focuses on
the transformation permissibility of the expression. Thus, his model is concerned with the semantic
properties of idioms, such as ‘frozenness’ and ‘analysability’. He placed idioms in a spectrum of various
levels which are based on the transformation potential of the idioms, rather than the lack of
transformation or productivity, as in Weinreich’s view.

Non-traditional approaches to idioms

The phrase logical approach

Various influential accounts into idioms have been proposed from phrase logical perspectives (e.g.
Cowie, 2001; Granger and Meunier 2008; Gries, 2008; Moon, 1998, Sinclair, 1991).Many scholars now
agree that a lexicon is not just a repository of single words but rather a dynamic system, which includes
‘larger lexical items’ (e.g.
Bybee, 2010; Croft and Cruse, 2004; Langacker, 1987). Phraseology is the field of research that is
largely concerned with these larger lexical items (i.e. formulaic sequences at large), and the context in
which a phrase is used, as well as its various usages. It is closely and directly related to lexicography and
builds on corpus linguistics. Gries (2008) observes that the interest in phraseology has grown
considerably over the last twenty years or so. In the past, phraseology was caricatured as “idiom
researchers and lexicographers classifying and researching various kinds of fairly frozen idiomatic
expressions”. In contrast, now it is concerned with theoretical research and practical applications which
have much more profound influence on researchers in the subdisciplines of linguistics, as well as in
language learning, acquisition, teaching and natural language processing (Gries, 2008, p.3).

Sinclair’s (1991) work represents an influential neo-Firthian account, which according to Malmkjær
(2004), places emphasis on the syntagmatic relations of lexis and runs in parallel with traditional
approaches to idioms. According to McEnery and Hardie (2012), a common trend in the neo-Firthian
approach is to view a word, its phraseology and collocational features as the keystones of linguistic
description. After an examination of the way texts are organized, Sinclair (1991) offered an explication
of how idioms operate in a given language. He proposed two principles that govern a speaker's choices
in sentence construction: (1) the open choice and (2) the idiom principle. The open choice principle
refers to the many options a speaker has to produce sentences in accordance with conventional
grammatical restrictions. For Sinclair, the open-choice principle reflects the normal way of describing
language, as it covers word-for-word operations. Formulaic language can be seen as a true embodiment
of Sinclair’s idiom principle and this view constitutes the core of linguistic structure. It entails word
choices not being random, and a speaker having at his/her disposal a large number of ‘preconstructed’
phrases, such as of course, which itself constitutes a single choice, even though it can be analysed as
segments. These two principles are complementary and work together in language production (cf. Wray,
2002). Put simply, the speaker's choice is not random; language users have freedom to choose, but are
constrained by the topic, context and register of the situation in language production, and therefore need
to resort to prefabricated items, including idioms.
Within this approach, Moon (1998) has also presented an influential model to explore what she called
Fixed Expressions and Idioms (FEIs). Her corpus findings report that almost 40% of English FEIs allow
lexical variation or transformation, and a smaller percentage has two or more variations. She
differentiates fixed expressions, a term that is similar to Weinreich’s stable collocations, from idioms.
She believes the word ‘fixed’ is not applicable to idioms, which she defines in terms of characteristics
such as compositionality and institutionalization where idioms can be placed along a scale of gradient
features. Scholarships from within the phraseological approaches share similar assumptions held in the
cognitive approaches− as will be presented in the next section, but they are concerned in particular with
multi-word units at large. Gries (2008) argues that cognitive grammar, as outlined in Langacker’s work
(1987) and Goldberg’s (1995) construction grammar are similar in their assumptions about
phraseologisms, although terminology differs. The difference he notes pertains to non-compositionality,
which he mentions when defining the parameters of phraseologisms.

The cognitive approaches

Cognitive linguists reject the idea that the mind has any unique and autonomous language-acquisition
device. This position is at odds with the generative grammar approach (e.g. Chomsky, 1980). Although
cognitive linguists do not necessarily disagree that part of human linguistic ability is innate and
autonomous, they deny that it is separate or different from non-linguistic processes (e.g. retrieval of an
event from memory). Cognitive linguistics also postulates that knowledge of language emerges from
language use, which supports calls to account for language use in the real world, as language does not
exist in a vacuum. In usage-based approaches, the emphasis is on the (dynamic) organization of
linguistic knowledge, which consists of structured networks of ‘constructions’ as defined by
construction grammarians (e.g. Fillmore et al., 1988) or the ‘schemas’ (mental representations) used by
cognitive grammarians (e.g. Langacker, 1987).

Tomasello (2003) adopting a usage-based cognitive linguistic perspective, regards language use as a
crucial aspect of the enculturation process and criticizes generative approaches. Tomasello questioned
the idea of the ‘innate grammar’ held in generative grammar. He is particularly critical of the concept of
‘grammatical sentences’ being the prime focus of Chomsky, at the expense of supposedly ‘non-
grammatical’ sentences which may actually be naturally encountered in everyday life such as idioms. He
therefore proposes an alternative usage-based cognitive theory, with accounting for language use and
social-pragmatic aspects also being central to his model. In construction grammar, the construction
represents the basic unit of grammar which is composed of symbolic configurations, defined as pairings
of form and meaning/function (e.g. Fillmore et al., 1988; Fillmore, 1988; Goldberg, 1995, 2006; Lakoff,
1987). Thus, a construction has to be meaningful. Construction grammarians hold that the term includes
phrases, idioms, single words and morphemes. The basic argument held in construction grammar is that
words and complex constructions (e.g. idiomatic expressions) are necessarily pairs of form and meaning,
regardless of their formal or internal semantic complexity.

Therefore, the merit of construction grammar as a holistic and (often) usage-based framework lies in its
commitment to treating all types of expressions equally. In the componential model adopted by the
generative approach (e.g. Chomsky, 1980), a single word is considered the largest unit of linguistic
components, which are governed by highly general rules of grammar. Lexicon and syntax are viewed as
separate, which means that some exceptions are not accounted for (e.g. idioms). Construction
grammarians, however, make no distinction between grammar and lexicon, and instead propose a
syntax-lexicon continuum.

Langacker’s theory of cognitive grammar is also constructionist in nature (cf. Langacker, 2005).
Cognitive grammar’s most fundamental assumption is that grammar is symbolic in nature; in
Langacker’s own words, "grammar is conceptualization” (Langacker, 1987, p. 494). In his view,
lexicon, morphology and syntax are not essentially different or separate. Langacker (1991) espouses a
lexicon-syntax continuum view, claiming that grammar resides in patterns to combine simpler symbolic
structures into more complex ones. In other words, any symbolically complex expression can, in his
view, be subsumed in the language, as long as it is symbolic, i.e. meaningful. Langacker (2008)
emphasizing the centrality of meaning in cognitive linguistics mentioned: If generative linguistics views
syntax as being central to language, Cognitive Linguistics accords this honor to meaning. The latter
seems far more natural from the perspective of language users. When ordinary people speak and listen, it
is not for the sheer pleasure of manipulating syntactic form- their concern is with the meaning expressed.
This does not of course imply that grammar is unimportant in language or in language teaching. It is,
however helpful to realize that grammar sub serves meaning rather than being an end itself. (p.67)

The assumptions held in cognitive grammar and cognitive linguistics are compatible with the
constructionist approach to idioms of Fillmore et al. (1988). Their notion of construction is essentially at
odds with the basic tenets of the generative model. The absence of the notion of constructions or any
similar assumption in the generative approach makes it impossible for this model to account for a
category in the language that cannot be simply captured by the ‘core grammar’, the ‘solution’ typically
being for these patterns to be relegated to the socalled ‘periphery’ (Chomsky, 1980).

Fillmore et al. in referring to the placement of idiomatic expressions whose meaning and use cannot be
deduced simply from knowledge of the constituent parts stated:

When constructions are interpreted as the products of maximally general rules, no place remains in the
grammar for spelling out the non-predictable semantics and pragmatics that is frequently conventionally
associated with particular constructions such as those we will describe [i.e. the the X-er, the Y-er
construction or the let alone construction]. (Fillmore et al., 1988, p. 507)

6. Conclusion

As has clearly been shown, linguistic approaches have gone beyond the idea of the number/type of items
that should qualify an expression to be considered an idiom. Scholars set criteria and identify
parameters, mostly semantic or syntactic when defining idioms. This of course is highly dependent on
the purpose of their analysis or niches which means that there is no one single definition for idioms.
Theoretically, Idioms have been treated differently in the traditional and non-traditional approaches.
Those approaches may seem opposing or different, which is partly true. They however benefit and build
on each other and are essentially complementary. Most of the criticism of the traditional, namely the
generative approaches to idioms were related to viewing idioms as a separate entity in the lexicon, and
relegating them to a secondary position in the language. Another criticism is that it ignores experience
and language use, and rather focuses on rule-governed constructs (which excludes some idioms).

Non-traditional approaches, particularly the cognitive linguistics approaches hold that lexicon and
grammar are not separate and propose a lexicon syntax continuum −, which is one of the principal
contrasts to the generative approaches. They also hold that language is a matter of use and conventions
irrespective of grammaticality. This view has resulted in a radical change in the cognitive-inspired
assumptions as represented in a myriad of recent theoretical and evidence-based studies. Particularly
applications on metaphor to which part of idioms were explained in terms of the idea of embodiment.
Moreover, some cognitive-based approaches argued that idioms are at least partly motivated and
associated metaphor to human cognition and experience. Therefore, such potentials can offer an
alternative framework to formal traditional theories of grammar and semantics to account for idioms at
the theoretical level and also empirically in studying the processing and comprehension of idioms in L1
and L2, as well as idiom teaching and learning.

Barcelona, A. (Ed.) (2003). Metaphor and metonymy at the crossroads: a cognitive perspective. Berlin:
Mouton de Gruyter.
Bybee, J. (2010). Language, Usage and Cognition. Cambridge: Cambridge University Press.
Carter, R. (1998). Vocabulary Applied Linguistics Perspectives, London, New York: Routledge.
Chomsky, N. (1965). Aspects of the theory of syntax. Cambridge, MA: MIT Press.
Chomsky, N. (1972). Studies on semantics in generative grammar. The Hague: Mouton.
Chomsky, N. (1980). Rules and Representations. New York: Columbia University Press.
Cieślicka, A. (2010). Formulaic language in L2: Storage, retrieval and production of idioms by second
language
learners. InM. Pütz & L. Sicola (Eds.) Cognitive Processing in Second Language Acquisition: Inside the
Learner’s Mind (pp. 149-68). Amsterdam, Philadelphia: John Benjamins.
Cieślicka, A. (2015). Idiom acquisition and processing by second/foreign language learners. In: Heredia,
R.R., &
Cieślicka, A. (Eds.) Bilingual figurative language processing. (pp. 208-244). New York:
Cambridge University Press.
Cowie, A. P. (Ed.) (2001). Phraseology: Theory, analysis and applications. Oxford: Clarendon Press.
Croft, W., & Cruse, D. A. (2004). Cognitive linguistics. Cambridge: Cambridge University Press.
Ellis, N. C. (2001). Frequency effects in language processing: A review with implications for theories of
implicit and explicit language acquisition. Studies in Second Language Acquisition, 24, 143–188.
Evans, V., & Green, M. (2006). Cognitive Linguistics: An Introduction. Edinburgh: Edinburgh
University Press. Fernando, C. (1996). Idioms and Idiomaticity. Oxford: Oxford University Press.
Fillmore, C. J. (1988). The Mechanisms of Construction Grammar.BLS, 14, 35-55. in Social Context (pp.
17-38). Amsterdam: North-Holland.
Fillmore, C. J., Kay, P. & O’Connor, M. K. (1988). Regularity and Idiomaticity: The case of let alone.
Language, 64, 501-538.
Firth, J. R. (1957). Modes of meaning .In Papers in linguistics 1934-1951 (pp. 190-215). Oxford:
Oxford University Press.
Langacker, R. W. (1987). Foundations of cognitive grammar: theoretical prerequisites (Vol. I).
Stanford, CA: Stanford University Press.
Langacker, R.W. (2005). Construction Grammars: Cognitive, radical, and less so. In F.J.R.D.M. Ibanez
&M.S.P.
Cervel (Eds.), Cognitive Linguistics: Internal Dynamics and Interdisciplinary Interaction (pp.
101-159). Berlin: Mouton de Gruyter.
Langacker, R. W. (2008). Cognitive Grammar as a Basis for Language Instruction. In P. Robinson and
N. C. Ellis (Eds.), Handbook of Cognitive Linguistics and Second Language Acquisition (pp. 66-
88). New York/London: Routledge.
Makkai, A. (1972). Idiom Structure in English. The Hague: Mouton.
Malmkjaer, K. (2004). The linguistics encyclopedia. Routledge: London.
McCarthy, M. (1998). Spoken language and applied linguistics. Cambridge: Cambridge University
Press.
McEnery, T., &Hardie, A. (2012). Corpus linguistics: method, theory and practice. Cambridge:
Cambridge University Press.
Mel’cuk, I. (1995). Phrasemes in language and phraseology in linguistics. In Everaert et al. (Eds.),
Idioms (pp.
167-232). Hillsdale N.J.: Lawrence Erlbaum Associates.
Moon, R. (1998).Fixed expressions and text: a study of the distribution and textual behaviour of fixed
expressions in English. Oxford Studies in Lexicology and Lexicography. Oxford: Clarendon
Press.
Newmeyer, F. J. (1972). The Insertion of Idioms.CLS, 8, 294-302.
Nunberg, G., Sag, I., & Wasow, T. (1994). Idioms. Language, 70, 491-538.
Widdowson, H. G. (2003) Defining issues in English language teaching. Oxford: Oxford University
Press.
Wray, A. (2002). Formulaic language and the lexicon. Cambridge: Cambridge University Press.
Wray, A. (2008). Formulaic Language: Pushing the Boundaries. Oxford: Oxford University Press.
Wulff, S. (2010).Rethinking Idiomaticity. Corpus and Discourse. New York: Continuum.
Zyzik, E. (2011). Second language idiom learning: The effects of lexical knowledge and pedagogical
sequencing. Language Teaching Research, 15, (4) 413-433.