To begin with, section 2 outlines the notion of idioms, distinguishes four different dimensions of idioms and explores the extent of grammatical and syntactic restrictions on idioms. Idioms are specific sequences of words whose meaning cannot be predicted on the basic meanings of its constituents. There are four different dimensions of idioms: encoding and decoding, grammatical and extragrammatical, substantive and formal, and finally, idioms with and without pragmatic point. Lexical idioms are grammatically and syntactically constrained due to the fact that they do not act as a grammatically single unit even though semantically they are one. In the section 3, one can find the interpretations of the concept of Construction Grammar and the notion of constructions.
Construction Grammar is a cognitive linguistic framework which explores and illustrates all grammatical knowledge in the form of constructions. Additionally, main distinctions between the Constructionist approaches and the mainstream Generative Grammar are listed and described. Constructions are combinations of form and function (semantic or discourse) which can include morphemes, words, idioms and more complex phrasal patterns.
Moreover, section 4 represents the central part of the work. Firstly, the connection between idioms and Construction Grammar has been made. It is argued that the Constructionist approach has emerged due to the necessity of finding an appropriate place for idioms within the language. Secondly, a short analysis of constructional idioms according to Booij and Jackendoff is given. Constructional idioms are considered to be a special type of construction with a partially or entirely non-compositional meaning in which the likely empty subset of the terminal components is established. They are further explored within the Verb Phrase. Additionally, three different case studies are described in order to present idioms as constructions. The case studies in question are the let alone case explored by Fillmore, Kay and OConnor, the way and ditransitive constructions by Adele Goldberg. All three of these cases are analysed within the constructional approaches as constructions.
2.1. Defining idioms
To begin with, an idiom is a sequence of words whose meaning cannot be predicted from the
Furthermore, Nunberg, Sag and Wasow claim that idioms have certain properties conventionality, inflexibility, figuration, proverbiality, informality and affect. Besides conventionality, none of the properties mentioned above applies obligatorily to all idioms. In other words, idioms are conventionalized in a sense that their meaning cannot be (entirely) predicted on the knowledge of the conventional use of their components when appear isolated from one another. Inflexibility refers to the limited number of syntactic constructions in which idioms appear while figuration includes metaphors, metonymies, hyperboles or other kinds of figurative meaning that idioms can typically involve. Regarding proverbiality, idioms can be used to describe a situation of particular social interest. This can be connected to informality which associates idioms with informal, colloquial and popular register. Finally, affect refers to the fact that idioms are typically used to imply a certain evaluation or affective stance towards the things they denote (Nunberg, Sag, Wasow 1994, pp. 492-3).
2.2. Dimensions of idioms
Fillmore et al. explore the dimension of idiomaticity and differ four different dimensions of idioms. Referring to Makkai (1972) they distinguish encoding and decoding idioms. Idioms of decoding are phrases that the language speakers are unable to completely and entirely accurately interpret without previously learning it. On the contrary, idioms of encoding are expressions that language users may or may not be able to comprehend without the preceding knowledge or use. A special characteristic of decoding and encoding idioms is that all decoding idioms are also encoding, but not all encoding are decoding idioms. For instance, expressions spill the beans, beat around the bush and costs an arm and a leg are decoding as well as encoding, but phrases such as perform surgery, heavy smoker and mindless chatter are instances of encoding idioms solely.
The second type of idioms these authors present are grammatical and extragrammatical idioms. To put is simply, some idioms consist of the words filling proper and familiar grammatical structures while some have words occurring in constructions which the rest of the grammar cannot account for (Fillmore, Kay, OConnor 1988, p. 505). For example, in grammatical idioms like pull someones leg, let the cat out of the bag or cook someones goose verbs and NPs are arranged in the exact manner one would expect them to be arranged. Conversely, there are anomalously structured expression such as first off, long live the king, long time no see, sight unseen and by and large which belong to the category of extragrammatical idioms. Their particular feature is the mentioned anomalous structure which is not made by following familiar rules of grammar and general application of those rules.
Thirdly, idioms can further be divided into substantive and formal, or in other words lexically filled and open idioms. Substantive or lexically filled idioms are described by more or less fully specified lexical make-up. Examples of substantive idioms are all of the idioms mentioned so far. On the contrary, formal or lexically open idioms raise some serious theoretical issues since their syntactic sequence has a certain semantic and pragmatic function unidentifiable from their form alone. Nevertheless, the fact that formal idioms can sometimes play the role of a host to substantive blurs the difference between the two idiom types in question.
One should compare the sentences (1) and (2). (1) The more carefully you plan your schedule, the easier it will be to get everything done. (2) The faster they run, the slower we run. Clearly, the first sentence is an innovative formation of the discussed syntactic sequence while the second sentence is fixed expression that uses the same form. Finally, one more distinction that should be made is the one between the idioms with and without pragmatic point. Idiomatic expressions can have a specific pragmatic purpose, e.g. a large number of substantive idioms (Good morning, See you later or once upon a time) have an obvious pragmatic function, but there is abundance of those which perform a contextually impartial and neutral function (like you cant have it both ways, either way is fine, etc.).
Furthermore, the the X-er the Y-er type of formal idioms is more or less liberated of pragmatic obligations as opposed to Him a doctor? type which has a particular pragmatic or rhetorical purpose (Fillmore, Kay, OConnor 1988, pp. 504-6).
2.3. Grammatical and syntactic restrictions on idioms
In section 2.1., it was mentioned that idioms do not act like single grammatical units even though concerning semantics they are one (Palmer 1976, p. 41). The reason for this is that there are many grammatical restrictions. For example, if the idiom to kick the bucket is seen as a grammatically single unit, then the past tense would be *kick the bucketed which, of course, is incorrect. Conversely, idioms do function to some degree as a normal sequence of grammatical words. Thus, the past tense of the idiom to kick the bucket is kicked the bucket. Moreover, although the verb within the idiom can be placed in the past tense, the number of the noun can never change. Therefore, one can say spilled the beans, but not *spill the bean.
Correspondingly, in the case of the idiom red herring, the noun can become plural (red herrings), but the adjective cannot be put into comparative (the -er) form. As a result, *redder herring may not be used. Regarding the syntactic restrictions, Palmer states that they are also very common among idioms. Namely, some idioms have passive while others do not. If an example The beans have been spilled is taken into consideration, it is obvious that this idiom can be used in passive voice, but to say *The bucket was kicked is not syntactically acceptable (Palmer 1976, p. 98).
What is more, elements of an idiom cannot take contrastive stress, or become the focus of topicalizing transformations (It was her brother’s leg that she pulled, not *It was her brother’s leg that she pulled) as well as they cannot be referred back to anaphorically (e.g. *Mary pulled her brother’s leg; John pulled it too). Also, it is not possible to substitute any of the idiom’s constituent elements by a synonym or near synonym (*The poor old chap kicked the pail) or to coordinate new elements with the original semantic constituents as in *She pulled and twisted her brother’s leg (Cruse 2011, pp. 86-7). Considering the latter cases, it can be concluded that some idioms are more restricted or frozen than others (Palmer 1976, p. 98). According to Curse, frozen metaphors are metaphorical expressions whose meaning and form have become relatively rigid (2011 p. 90).
3. The concept of Construction Grammar and the Constructionist Approaches
To begin with, Construction Grammar is a cognitive linguistic approach to syntax and a grammatical scheme which essentially presents the entire knowledge of language rules equally, in the form of constructions. The notion of Construction Grammar arose during the time between 1960s and 1980s as a response to the grammatical knowledge of the generative grammar (Croft, Cruse 2004 p. 225). Moreover, it is widely known that the Saussurean notion of the linguistic sign implies an arbitrary and conventional pairing of form (or sound pattern/signifiant) and meaning (or mental concept/signife) (Hoffman and Trousdale 2013, p. 1).
Over more than seventy years after his death, some linguists began to investigate the possibility of arbitrary form-meaning pairings being used not exclusively for illustrating words and morphemes but also for describing all levels of grammar (Hoffman and Trousdale 2013, p. 1). This kind of usage of the Saussurean sign developed a particular linguistic approach referred to as Construction Grammar whose main focus lies in the notion of construction that contains different linguistic units such as morphemes, words, idioms, phrasal patterns etc. Considering all constructions to be a part of a lexicon-syntax continuum, Construction Grammarians did not draw a clear division between lexicon and syntax in their framework unlike Traditional Generative Grammarians have (Hoffman and Trousdale 2013, p. 1).
Basically, the revival of Construction Grammar happened as a reaction to the componential model of the organization of a grammar that is found in generative syntactic theories (Croft 2007, pp. 463). In the componential model all characteristics of an utterance are explained within several different components phonological, syntactic or semantic. This means that each component provides and governs linguistic properties of a single type. The only exception are words because they contain phonological form, syntactic category, meaning and therefore are essential for the existence of an independent component lexicon (Croft 2007, pp. 464-5). Simply, it can be said that the lexicon is the vertical constituent in opposition to the horizontal elements which can be syntactic, semantic or phonological (Croft, Cruse 2004, p. 226). Furthermore, the componential model offers the presentation of grammatical knowledge without referring to the constructions. Consequently, the logical conclusion of componential analysis would be that all grammatical structures larger than one word can be explained by the use of general rules of the previously mentioned components (Croft 2007, pp. 464-5).
On one hand, these two approaches are aware of the importance to perceive the language as cognitively organized system, accept the fact that combining structures in order to create novel utterances are required, as well as acknowledging the necessity for non-trivial theory of language. On the other hand, constructionist approaches, unlike mainstream generative grammar, tend to highlight the distribution and detailed semantics of certain grammatical morphemes, words and cross-linguistically uncommon phrasal patterns (Goldberg 2006, pp. 45).
Another crucial discrepancy between generative and constructional approaches is that the mainstream generative grammar argues the language speakers are required to be well acquainted with the language faculty doctrine or in other words, they must be familiar with the universal grammar. The fact is that the core phenomena is much easier to acquire due to its regularity and frequency. Generative approaches placed idiosyncratic and semi-idiosyncratic constructions on the periphery of language whereas constructionist approaches are arguing that whatever means we use to learn these patterns can easily be extended to account for socalled core phenomena (Goldberg 2003, p. 222) There is a tendency among linguists to assume that these possibly confusing and primarily peripheral cases should be acquired on the grounds of previously obtained language knowledge. Conversely, theories of Construction Grammar suggest that more regular, general and commonly repeated cases could also be learned in the same way. Another difference provided by Goldberg, in opposition to mainstream generative theories, according to the Constructionist approaches, one construction does not derive from another (Goldberg 2003, pp. 221-2).
In addition, Croft and Cruse present further characteristics of Construction Grammar as well as the differences between the Construction Grammar and componential syntactic theories. Firstly, the main difference between the two frameworks is that the symbolic link between form and conventional meaning is internal to a construction in the latter, but is external to the syntactic and semantic components in the former (Croft, Cruse 2007, p. 258) Secondly, in construction grammar, the basic linguistic units are symbolic and therefore organized as symbolic units while in the componential model, syntactic forms are arranged independently of the equivalent semantic forms. Structure of basic units in the construction grammar is more complicated. (Croft, Cruse 2007, pp. 256-0).