Date of Award


Publication Type


Degree Name



L. Buchanan


Adjective-nouns, Conceptual combination, Concreteness, Semantic neighborhood density, Semantic processing



Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.


Conceptual combination is an active meaning construction process involved in the production and comprehension of complex concepts (e.g., SLEEP TREE, STONY FACE). Distributional and schema-based theories of conceptual combination have proposed various cognitive mechanisms with a primary focus on the processing of noun-noun complex concepts (e.g., SLEEP TREE). The manipulation of variables related to the constituent (e.g., relational frequency) and phrase (e.g., typicality) composition has provided insightful advances into the conceptual representation and processing of complex concepts. Within this context, semantic variables related to semantic richness and concreteness of complex concepts have not been examined in the conceptual combination literature despite having been thoroughly investigated with respect to the processing of simple concepts.

The primary objective of the current study is to investigate the processing of adjective-noun combinations (e.g., STONY FACE) by manipulating semantic variables related to the constituent (i.e., semantic neighbourhood density or SND) and phrase (e.g., concreteness) structure. The adjective-noun stimulus set was constructed with participant ratings using a novel quantitative measure to capture a varying degree of novelty (Experiment 1a) and concreteness (Experiment 1b). In the remaining experiments, the processing of adjective-noun combinations was examined with methodology capturing online processing with tasks of differential semantic engagement (Experiments 2-4) as well as an offline interpretation task (Experiment 5). Collectively, the findings of the current study inform our understanding of the conceptual representation and comprehension of adjective-noun phrases. The results of the online processing experiments demonstrated orthographic and semantic effects, which were observed in a graded fashion based on the level of semantic processing the task required. In the shallowest double lexical decision task with non-pronounceable non-words (Experiment 2), only orthographic effects pertaining to the visual word form of adjective-noun phrases were found (i.e., combined letter length, mean orthographic frequency). In Experiment 3, where non-words were pronounceable and required a deeper level of semantic processing compared to Experiment 2, a partial meaningfulness effect was observed, as high meaningful adjective-noun pairs had faster response latencies compared to low meaningful adjective-noun pairs, though no differences were observed for the intermediate meaningful group. A concreteness effect, in which concrete word pairs are processed faster relative to abstract word pairs, was also observed in Experiment 3, particularly for low meaningful adjective-noun phrases. Complete main effects of meaningfulness and concreteness were observed in Experiment 4, the deepest semantic processing task that required participants to make judgments about whether adjective-noun pairs made sense as a pair, essentially recruiting conceptual combination under pressured time constraints. SND effects were also prominent in Experiment 4 and yielded asymmetrical modifier and noun effects based on the meaningfulness and concreteness of the phrase. In Experiment 5, participants were asked to provide an explicit interpretation of novel (low meaningful) adjective-noun phrases, and four themes of interpretation types were identified, including slot-filling, noun elaboration, abstraction, and adjective-reversal. The proportion of unique interpretations and interpretation types differed based on the semantic composition of the adjective-noun phrases.

The results were taken as further support for language-based models of conceptual representations, based on the SND effects observed in Experiment 4 and 5, as SND is a quantitative variable derived from a language-based co-occurrence model (Durda & Buchanan, 2008). Kintsch’s (2000) computational model of constructing sentence meaning was applied as a mechanism of constructing meaning for adjective-noun phrases using Experiment 4 and 5 findings, based on previous results in adjective-noun metaphors (Al-Azary et al., 2021). This model can account for a variety of points made by other theorists of conceptual combination, including recruitment in both familiar and novel phrases, an important role of the modifier, an interaction between modifier and noun constituents, competition among different potential processing routes, and recruitment of prior background knowledge.

Included in

Psychology Commons