Project members:
Simone Löhndorf, Peter Gärdenfors, Carita Paradis, Kristina Hansson, Rasmus Bååth, Jens Larsson.

Aim of project/main objective/main hypothesis:

The goal is to study the semantic development of adjectives in children and adolescents. In particular the degree of abstract uses of adjectives will be investigated, i.e. the trajectory from concrete language use to the use of more abstract language and metaphors and metonymies in later language development. The nature of the noun and the salience of the aspect of the meaning of the noun that is modified by the adjective is important for the interpretation of adjectives. The adjective–noun combination is examined as a whole, as well as in its parts. A well-known problem with adjectives is that they are not possible to study in isolation; one has to take the noun into account.  A multiple factor analysis will be carried out, identifying the abstractness of the nouns and what kind of information the adjectives target in the noun. The focus is on word sense from the perspective of cognitive semantics, where lexical meanings are considered to be dynamic and sensitive to contextual demands, rather than fixed and stable.

Currently the model of Ontologies and Construals (Paradis 2005) is used for the tagging of the nouns and adjectives. Paradis distinguishes three types of nominal content structures: first-order meaning structures, second-order meaning structures and third-order meaning structures. First order meaning structures are physical objects that exist in time and space, such as for example PEOPLE, ANIMALS, and ARTEFACTS. Second order meaning structures comprise EVENTS, PROCESSES/ACTIVITIES and STATES, i.e. things that exist in time. Third-order meaning structures are represented as ABSTRACT meanings. Like nouns, adjectives profile both first-order meaning structures, second-order meaning structures and third-order meaning structures. However, most often adjectives are second order meaning structures, namely STATES, expressing simple properties such as ‘a thin man’ or ‘a good book’. This is a corpus based approach to study the developmental path of the production and various usages of adjective-noun combinations, from the age of 10 to adult language use.

Gärdenfors’s theory about the domain-specifity of meaning leads to the prediction that words are not learnt one buy one but in domain. So when one color word is learnt by a child, others will follow soon. In particular the implies that when a polar adjective is learnt, its opposite will be learnt at about the same time in linguistic developoment. This prediction has been tested on 3 and 5 year old children by Siri de Geer and Cristina Gunnarsdottir. Their results support the theory of developing domains for the learning of adjectives.

Why is this project interesting?
The development of the semantics of adjectives has hardly been studied. The use of adjectives is scarce, i.e. children talk about different objects and phenomena, but they don’t describe or specify them further. With increasing age, the knowledge of word meanings goes beyond the most concrete sense and words are used in different senses in a variety of contexts. Furthermore concrete words are used metaphorically to express abstract meanings, and vice versa, abstract words may by way of reification, be used to denote concrete meanings. But when does this happen? And what kind of abstract meanings do emerge first and in what contexts? Knowledge about when and in what ways the meanings of adjectives extend over time to become more abstract can give us more general insight into conceptual development and deviations from typical development.
The test concerning learning opposites of adjectives that was tested by de Geer and Gunnarsdottir can hopefully be extended to a test that can be used clinically to identify children with non-normal semantic development. This would ba a prime example of how a semantic theory directly generates a clinical evaluation test.

Potential applications (diagnostic tools, methods for other research areas, for society):

The fairly large Swedish text corpus that has been generated will have great potential for other researchers. It is possible to perform a large number of linguistic analyses on the texts in order to find out more about language development, how children structure their texts, etc. Since the teachers grades are available, research could also be conducted on which criteria teachers use to set different grades. If the experimental, psycholinguistic study, provides us clues about how concrete and abstract concepts are processed in the brain at different ages, the results could provide guidelines about what linguistic material it is reasonable to expect children at different ages to understand and use.

Conferences, workshops, lectures, invited talks:

S. Löhndorf: Poster presentation at the Association for Research and Applying Metaphor (RaAM) conference, Almagro, May 4-7, 2011:
S. De Geer and C. Gunnarsdottir: “The acquisition of adjective antonyms in Swedish-speaking children”, paper presented at the Workshop on the Acquisition of Adjectives Across Languages, Utrecht, 28-29 November 2013.

Collaborations within CCL

Birgitta Sahlén, Kristina Hansson, Sverker Sikström on semantic maturity in narrative.

Magnus Lindgren for a joint psycholinguistic ERP-study.


Victoria Johansson on writing development.

Roger Johansson on mental imagery of concrete and abstract concepts.

Olof Sandgren on adjective usage of children with language impairment.

Collaborations outside of CCL

Carita Paradis, Department of English. Peter is developing an international network with researchers on formal models of semantics that can be invited as guest lecturers/researchers to CCL.