UFJF - Universidade Federal de Juiz de Fora

What’s in a word? – Item-specific and general knowledge in linguistic theory

You are at: Home > Theme sessions > What’s in a word? - Item-specific and general knowledge in linguistic theory


Peter Uhrig | FAU Erlangen-Nürnberg


It is commonplace for linguistic theories to assume that there is general knowledge and that there is item-specific knowledge in language. Traditionally, the grammar is the locus of the general (often released) knowledge whereas the lexicon is where the item-specific information is stored (e.g. Sweet 1899/1964, Bloomfield 1933/1935, Quirk et al. 1985, Pinker 1999).

Challenges to such a dichotomy, which also was a primary tenet of generative grammar (e.g. Chomsky 1965, 1981), came from several camps:

–  Corpus linguistics with researchers such as Sinclair, whose Idiom Principle2  (Sinclair 1991) and whose take on lexical items (Sinclair 1998) dissolve the boundaries between lexis and grammar.

–  Psycholinguistic studies on various aspects (e.g. Bybee 1995, 2007, 2010), including language acquisition (e.g. Tomasello 2003, Lieven/Behrens/Speares/Tomasello 2003), that produce evidence against a view of language use that combines atomic lexicon entries with the help of abstract rules.

–  Cognitive approaches to grammar, for instance Cognitive Grammar (e.g. Langacker 1987, 1999), where “[l]exicon, morphology, and syntax form a gradation claimed to be fully describable as assemblies of symbolic structures” with “degrees of schematicity” (Langacker 1999: 122), or Construction Grammar (e.g. Fillmore/Kay/O’Connor 1988, Goldberg 1995, 2006), where the distinction between lexicon and grammar is also given up.

Although the distinction between lexicon and grammar is given up in constructionist approaches, most researchers agree that our stored knowledge forms a continuum between concrete items, such as words and their properties (e.g. grammatical gender in German or French), and more abstract linguistic units, such as Argument Structure Constructions in the sense of Goldberg, or even constituent ordering according to weight and information structure. What is controversial for grammarians is the extent to which storage of item-specific information takes place and is responsible for language use. Thus a valency grammarian may treat him  in She baked him a cake.  as an optional complement of bake  while a Goldbergian constructionist would prefer an analysis in which him  is supplied by the construction and not by the argument list of the verb. 

There are, however, further questions related to the storage of linguistic knowledge. One pertains to the fact that during language acquisition the constructicon is a set of many “construction sites” with the stored knowledge being constantly updated, re-arranged and re-interpreted. What exactly this means for the form in which we store these items has not been fully understood yet. Holistic processing is a domain-general cognitive process, which is fully compatible with her view that memory for language is similar to memory for other things (e.g. events, sequences, motor activity).

Keywords: item-specificity, generalizations, constructicon, storage


Accepted Papers:

What is item-specific and what is general in lexical knowledge in which processes?
Heike Behrens

Holistic Processing: Effect on items and constructions.
Joan Bybee

New verbs of communication in English and German from a constructionist and a valency angle
Adele Goldberg & Thomas Herbst,

Continuous frame/content shaped interchangeably in idiosyncratic verb and tense element

Edmond Cane

Telicity and verb meaning in the choice of Transitive or Intransitive motion constructions in Brazilian Portuguese

Cida Oliveira

The particle out in the multi-word verb come out

Edelvais Barbosa

Metaphorical extensions of OVER: a study on Functional-cognitive linguistics

Raquel Rossini