The number meaning of grammatically plural nouns is to some extent context sensitive. In negative sentences, plural nouns typically receive an inclusive reading referring to any number of individuals (one or many). This contrasts with their more frequent exclusive reading referring to a group of two or more individuals. The present study investigated whether a plural noun in a negative sentence is treated as inclusive immediately when it is encountered or whether this interpretation is delayed. In an experiment using a technique based on a numerical variant of the Stroop effect (Berent et al. in J Mem Lang 53:342–358, 2005. 10.1016/j.jml.2005.05.002; Patson and Warren in J Exp Psychol Learn Mem Cogn 36(3):782–789, 2010. 10.1037/a0018783), participants counted visually presented singular and plural Polish nouns embedded in either affirmative or negative sentences. The nouns were displayed once or as two copies. Plural nouns were easier to count when they were repeated twice on the screen than when only one copy was displayed. For singular nouns this pattern was reversed and the effect was weaker. Crucially, no difference was found for plural nouns appearing in affirmative and negative sentences. This indicated that an inclusive (“one or more”) reading of plural nouns in the scope of sentential negation was not immediate. The results are in line with past research suggesting that the semantic processing of a negative sentence may proceed in two phases (Fischler et al. in Psychophysiology 20(4):400–409, 1983. 10.1111/j.1469-8986.1983.tb00920.x; Kaup et al. in J Pragmat 38:1033–1050, 2006. 10.1016/j.pragma.2005.09.012; Lüdtke et al. in J Cogn Neurosci 20(8):1355–1370, 2008. 10.1162/jocn.2008.20093; Spychalska in Proceedings of the 2011 ESSLLI student session, 2011).

The article is a contribution to a long-standing discussion on how idioms are represented and accessed in the mental lexicon. More specifically, in a timed cloze response study we investigate high and low syntactic flexibility idioms in Polish in order to find out whether the degree of syntactic flexibility influences the ease and time of idioms’ recognition. By doing so we contribute to the question of whether idioms are differently represented in the lexicon depending on their syntactic flexibility, as suggested by Nunberg et al. 1994 and Gibbs and Nayak 1989, or whether all idioms independent of their syntactic flexibility are represented lexically in the same hybrid way, as suggested by Cutting and Bock 1997 and Sprenger et al. 2006. The results of our study support the latter view.

The current paper presents results of two experiments attempting to replicate with Polish speakers a Stroop-like interference of grammatical number with the counting task, first reported by Berent et al. (2005) for Hebrew. Both experiments tested the influence of the type of number morphology (marked with overt suffix vs. unmarked) of nouns on the strength of the interference effect. Additionally, the second experiment investigated the processing of nouns with a mismatch between grammatical and conceptual number and tested the possible effect of animacy on number interpretation in order to determine the time at which the information about grammatical number is activated. The first experiment showed a significant interaction between the grammatical number and visual numerosity of the counted words and the effect of markedness, with marked singulars producing a bigger congruency effect than unmarked singulars. However, in the second experiment the influence of morphology was reversed and the overall effects were considerably weaker.

No abstract

Summary This article reports on two timed cloze-response experiments which examine the impact of context on idiom recognition. Study 1 presented participants with the beginnings of Polish VP idioms without any prior context. Cloze probabilities and response times for idiom continuations were measured to establish the idiom recognition point (IRP) for each idiom. In Study 2, we used the same idioms in two kinds of contexts: (i) supporting a figurative meaning and (ii) supporting a literal meaning. Cloze probability and response times were measured at the IRP and one word before and after it. The figurative meaning of idioms was automatically activated at the IRP independently of the type of context. Additionally, the figurative context did not move the IRP to an earlier position, whereas in the literal context the responses were significantly slower at the IRP as compared to the figurative context condition. Such a finding indicates that, irrespective of the literal context, the comprehenders automatically activated the figurative meaning of an idiom at the IRP, but they had to discard it later. The literal meaning was computed from the literal meanings of idiom constituents stored in idiom lexical representation, which was computationally costly.