High-frequency words are processed faster than low-frequency words, known as the word frequency effect (FE). Although the FE has been studied in various writing systems as well as in first- (L1) and second-language (L2) reading, existing theoretical hypotheses are mainly based on findings in alphabetic languages. To date, no study has investigated theoretical explanations of the FE such as the learning hypothesis, the lexical entrenchment hypothesis and the rank hypothesis apply to Chinese–English bilinguals. The present study, therefore, compared the FEs in Chinese– and Dutch–English bilinguals during natural paragraph reading in their L1 and L2, using eye-tracking measures. Chinese bilinguals exhibited a larger FE in L2 than in L1. They displayed smaller L1 FEs and much steeper L2 FE curves than Dutch bilinguals. These findings are not entirely consistent with the existing FE hypotheses, and the present study discusses theoretical accounts in light of the observed results.

We examined how noun frequency and the typicality of surrounding linguistic context contribute to children’s real-time comprehension. Monolingual English-learning toddlers viewed pairs of pictures while hearing sentences with typical or atypical sentence frames (Look at the… vs. Examine the…), followed by nouns that were higher- or lower-frequency labels for a referent (horse vs. pony). Toddlers showed no significant differences in comprehension of nouns in typical and atypical sentence frames. However, they were less accurate in recognizing lower-frequency nouns, particularly among toddlers with smaller vocabularies. We conclude that toddlers can recognize nouns in diverse sentence contexts, but their representations develop gradually.

This study investigates the acquisition of the Imperfective verb inflection paradigm in Emirati Arabic (EA), to determine whether the learning process is sensitive to the phonological and typological properties of the input. We collected data from 48 participants aged 2;7 to 5;9 years, using an elicited production paradigm. Input frequencies of inflectional contexts, verb types and tokens were obtained from corpora of child-directed and adult EA. Children's accuracy was inversely related to the input frequency of inflectional contexts, but not related to type and token frequency or phonological neighborhood density. Token frequency interacted with age, such that younger children performed considerably worse on low-frequency tokens, but older children performed equally well on high- and low-frequency tokens. We conclude that learning is input-driven, but that a sufficiently regular paradigm allows children to eventually generalise across all items earlier than in previously studied European languages.

Inflection impairments are commonly noted in aphasia, particularly non-fluent variants, where descriptions of such difficulties often focus on inflection omission. This aligns with rule-based theory, in which inflected forms should be more difficult to produce than their uninflected counterparts. Recent studies address noun inflection for number and potential effects of the relative frequency of singular and plural forms (dominance effects). However, none examine number errors qualitatively or in spontaneous speech. We present quantitative and qualitative analyses of such errors in nouns produced by twelve people with aphasia in spoken Cinderella narratives, examining: error rate; error types and nouns involved; relationship between error production and dominance; and speakers’ consistency with error production and flexibility in varying the noun forms concerned. Twelve unimpaired speakers provide comparison data. While error rates were low, arguably more important is error type. Singularisation and pluralisation errors were observed, all on regular nouns and involving production of the dominant form. The pluralisation errors, all occurring on references to Cinderella’s glass slipper, arguably challenge rule-based predictions that the singular is easier to retrieve than the plural. We suggest constructivist, usage-based theory as a promising framework to characterise such productions. Implications for aphasiology and clinical practice are also discussed.

Usage-based approaches to language acquisition posit that first (L1) and second language (L2) speakers should process more frequent compositional phrases, which have a meaning derivable from word parts, faster than less frequent ones (e.g., Bybee, 2010; Ellis, 2011). Although this prediction has received increasing empirical support, methodological limitations in previous relevant studies include a lack of control of frequencies of subparts of target phrases and scant attention to L2 production. Addressing these limitations, the current study tested phrase frequency effects in both language comprehension and production in two respective experiments, in which adult native English speakers (N = 51) and English L2 learners (N = 52) completed a timed phrasal decision task and an elicited oral production task. Experiment 1 revealed phrase frequency effects in both groups, lending support to usage-based researchers’ proposal that L1 and L2 speakers retain memory of word co-occurrences and that compositional phrase processing reflects an accumulation of statistics in previously encountered input. Experiment 2, however, provided weaker evidence for phrase frequency effects in these participant groups. Based on the results and previous empirical studies, methodological issues that may have impacted frequency effects and implications for future work in this area are discussed.

Phonological characteristics and frequencies of stems and allomorphs have been explored as possible factors causing differences in production accuracies between allomorphic forms. However, previous findings are not consistent and the relative contributions of these factors are unclear. This study investigated target and erroneous productions of the Dutch diminutive, which has five allomorphs with varying type frequencies and of which the selection depends on the phonological characteristics of the stems. Typically developing children (N = 115, 5;1–10;3) were tested on their production of real and nonce diminutives. Linear mixed effects modelling was used to analyse the data taking nonverbal IQ into account. Type frequencies of the allomorphs and differences in phonological characteristics of the stems were found to be related to differences in production accuracies between the allomorphs. However, phonological characteristics of the stems appeared to have a bigger impact, mainly due to the phonological complexity of these characteristics.

This paper tests whether lemma frequency impacts the duration of homographic noun–verb homophones in spontaneous speech, e.g. cut (n)/cut (v). In earlier research on effects of lemma frequency (e.g. Gahl 2008), these pairs of words were not investigated due to a focus on heterographic homophones. Theories of the mental lexicon in both linguistics and psycholinguistics differ as to whether these word pairs are assumed to have shared or separate lexical representations. An empirical analysis based on spontaneous speech from the Buckeye corpus (Pitt et al. 2007) yields the result that differences in lemma frequency affect the duration of the N/V pairs under investigation. First, this finding provides evidence for N/V pairs having separate representations and thus supports models of the mental lexicon in which lexical entries are specified for word class. Second, the result is at odds with an account of ‘full inheritance’ of frequency across homophones and consequently with speech production models implementing inheritance effects via a shared form representation for homophonous words. The findings are best accounted for in a model that assumes completely separate lexical representations for homophonous words.

‘Special reduction’ refers to instances of extreme phonetic reduction which is restricted to particular words or phrases, usually grammaticalizing constructions (going to > [gə̃ɾ̃ə̃]), greetings (hi from how are you), discourse markers (Spanish o sea > sa), or other sequences that are often used together. On the basis of data from English, Brazilian Portuguese, and Colombian Spanish, we argue that special reduction is based on the general phonetic tendencies in the language, but that these tendencies are carried to an extreme where word sequences are used with high frequency and become chunked, allowing formerly stressed syllables to lose stress and reduce. The data also show that special reduction takes place gradually over time, and reflects general patterns of change seen in the history of the language. In fact, in some examples, special reduction presages more general sound changes that occur later. We argue that the gradual phonetic changes that accumulate for particular words or phrases, eventually changing them dramatically, requires an exemplar model for the phonological representation of words and phrases, which is updated continually as sequences are used and affected by reductive phonetic processes.

The article presents a real-time study of standardization and regionalization processes with respect to the use of past participles of strong verbs in the western part of Denmark. Analyses of a large corpus of recordings of informants from two localities show that the use of the dialectal en form of the past participle suffix has been in decline during the last 30 years. The en forms are replaced by three other forms, one of which is (partly) dialectal, one regional and one standard Danish. The study indicates that a regionalization process has taken place prior to the time period studied, but that it has now been overtaken by a Copenhagen-based standardization process. The study also shows interesting differences between the two localities, arguably due to the geographical location and size, and to the status of the different participle forms in the traditional local dialects.