The Mental Lexicon - Volume 6, Issue 2, 2011

Research on several Indo-European languages attests to notable difficulties in inflecting irregular nouns and verbs. In these languages morphological and phonological factors are often intertwined in a way that obscures the source of the problem. Hebrew by contrast allows isolation of morphological and phonological factors in nominal inflection. Three experiments demonstrated that as in Indo-European languages, nominal inflection of Hebrew irregular nouns is slower than that of regular nouns and involves more errors. The occurrence of phonological alterations to the noun’s stem with the inflection is an additional source of irregularity, which also taxes the inflectional process in reaction time and error rate. The empirical results underline the power of the default automatic suffixation process as the main obstacle to irregular inflection. A theoretical contribution of this study is an interpretation of the irregularity effect based on a morphological analysis that views Hebrew as having a linear rather than a non-linear morphology. The stem–suffix match is suggested as the dominant factor affecting the inflectional process, responsible for the difficulties in irregular inflections. It is argued that in Hebrew, the differences between inflecting regular and irregular nouns can be easily and adequately explained as resulting from a mismatch between a stem and an affix.

Nonlinear methods of fractal analysis and recurrence quantification analysis are becoming more commonplace in the cognitive and behavioral sciences. These methods are illustrated here in a tutorial style using self-paced reading data. Self-paced reading was performed in which each spacebar press revealed a story word-by-word or else sentence-by-sentence. Participant readers were either Ph.D. candidates in English literature or undergraduates from an introductory psychology course and the same story was read by all, either one time only or reread another time on another occasion. The nonlinear analyses revealed crucial differences between the word unit and sentence unit conditions. Performance in the word unit condition was dominated by a task specific strategy, yielding data patterns more like those observed in tapping tasks. Nonlinear analyses of the sentence unit condition, however, discriminated between graduate and undergraduate readers, and first readings of the story from re-reading. From these analyses, the repeated reading of the same story reveals a kind of über-fluency, in a manner of speaking, of the Ph.D. candidates in English literature, whose performance stayed at or closer to a performance ceiling in both readings.

Most two-syllable English nouns exhibit trochaic stress, whereas most two-syllable English verbs exhibit iambic stress. This study examined whether ESL speakers whose native language is not a stress language acquire knowledge of stress typicality in English. Thirty native Chinese and 30 native English speakers participated in two grammatical decision tasks. Experiment 1 used English noun-verb homographs embedded in two grammatical frames (‘the ___’ and ‘to ___’). Both groups showed different response patterns to ungrammatical phrases. Experiment 2 used two-syllable non-homographs: typically-stressed nouns and verbs and atypically stressed nouns and verbs. Results indicated that stress typicality and grammatical frame both create expectations regarding word type, and these expectations affect lexical access in Chinese and English listeners in similar ways.

The present study investigates the processing and production of four-word sequences such as I don’t really know, at the age of, and I think it’s the. Specifically, we investigate the influence of families of probabilistic measures such as unigram, bigram, trigram, and quadgram frequency of occurrence, logarithmic (log) probability of occurrence, and mutual information. Log probability of occurrence emerged as the predominant predictor family in the onset latency analysis, suggesting that recognition is mainly underpinned by competition between a target N-gram and its family members. In contrast, the amount of experience one has with an N-gram (frequency of occurrence) surfaced as the most prominent predictor in production. Further, probabilistic measures tied to trigrams surfaced as the best predictors in the onset latency analysis, while the measures tied to unigrams were most predictive of production durations.Finally, the interactions between probabilistic measures tied to unigrams, bigrams, trigrams, and quadgrams suggest that N-grams of different lengths are processed in parallel in both recognition and production.

Barenboym, Wurm, and Cano (2010) recently showed that significant differences emerged for ratings gathered online and in person. They also showed that researchers could reach different statistical conclusions in a regression analysis, depending on whether the norms were gathered online or in person. In the current study that research was extended. Familiarity ratings gathered online were significantly higher than those gathered in the lab, for a set of 300 potential stimuli. The in-person ratings correlated significantly better with an existing database of familiarity values. It is also shown that under three different grouping methods, online and in-person familiarity ratings produce different sets of stimuli. Finally, it is demonstrated that in each case, different conclusions are reached about variables that have a significant relationship with familiarity. Simulations show that the effects are driven disproportionately by higher intra-item variability in the online ratings. Studies in which stimuli are grouped on the basis of ratings can be affected by the choice of rating methodology.