Content uploaded by Martin Corley
Author content
All content in this area was uploaded by Martin Corley on May 22, 2015
Content may be subject to copyright.
A preview of the PDF is not available
Disfluency in dialogue: An intentional signal from the speaker?
Abstract and Figures
Disfluency is a characteristic feature of spontaneous human speech, commonly seen as a consequence of problems with production. However, the question remains open as to why speakers are disfluent: Is it a mechanical by-product of planning difficulty, or do speakers use disfluency in dialogue to manage listeners' expectations? To address this question, we present two experiments investigating the production of disfluency in monologue and dialogue situations. Dialogue affected the linguistic choices made by participants, who aligned on referring expressions by choosing less frequent names for ambiguous images where those names had previously been mentioned. However, participants were no more disfluent in dialogue than in monologue situations, and the distribution of types of disfluency used remained constant. Our evidence rules out at least a straightforward interpretation of the view that disfluencies are an intentional signal in dialogue.
Figures - uploaded by Martin Corley
Author content
All figure content in this area was uploaded by Martin Corley
Content may be subject to copyright.
... The role of hesitations in human-human interaction is a much-discussed topic in current research [11,12], and there is evidence for the hypothesis that they have a communicative function and influence the interaction partner. As hesitations occur more often prior to long utterances [13] and infrequent words [14], the frequency distribution of hesitations in spontaneous speech is assumed to be caused by the level of cognitive load of the speaker. ...
... Through a shallow grid search, we obtained optimized results for all participants with a parameter setting of 500 trees, a maximum tree depth of 10, a minimum sample split of 3, and a minimum of 4 samples per leaf. Further, a combination of all features from the delta (1-4 Hz), theta (4-8 Hz), alpha (8)(9)(10)(11)(12), low beta (12)(13)(14)(15)(16)(17)(18)(19)(20), and high beta (20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30) bands binned in 2 Hz led to the optimal performance for 4-12 Hz in the first classification task (ambient vs. distraction) and for 4-20 Hz in the second classification (distraction vs. hesitation), and 62 × 5 = 310 and 62 × 9 = 558 dimensional feature vectors, respectively. ...
... Through a shallow grid search, we obtained optimized results for all participants with a parameter setting of 500 trees, a maximum tree depth of 10, a minimum sample split of 3, and a minimum of 4 samples per leaf. Further, a combination of all features from the delta (1-4 Hz), theta (4-8 Hz), alpha (8)(9)(10)(11)(12), low beta (12)(13)(14)(15)(16)(17)(18)(19)(20), and high beta (20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30) bands binned in 2 Hz led to the optimal performance for 4-12 Hz in the first classification task (ambient vs. distraction) and for 4-20 Hz in the second classification (distraction vs. hesitation), and 62 × 5 = 310 and 62 × 9 = 558 dimensional feature vectors, respectively. ...
In this paper, we investigate the effect of distractions and hesitations as a scaffolding strategy. Recent research points to the potential beneficial effects of a speaker’s hesitations on the listeners’ comprehension of utterances, although results from studies on this issue indicate that humans do not make strategic use of them. The role of hesitations and their communicative function in human-human interaction is a much-discussed topic in current research. To better understand the underlying cognitive processes, we developed a human–robot interaction (HRI) setup that allows the measurement of the electroencephalogram (EEG) signals of a human participant while interacting with a robot. We thereby address the research question of whether we find effects on single-trial EEG based on the distraction and the corresponding robot’s hesitation scaffolding strategy. To carry out the experiments, we leverage our LabLinking method, which enables interdisciplinary joint research between remote labs. This study could not have been conducted without LabLinking, as the two involved labs needed to combine their individual expertise and equipment to achieve the goal together. The results of our study indicate that the EEG correlates in the distracted condition are different from the baseline condition without distractions. Furthermore, we could differentiate the EEG correlates of distraction with and without a hesitation scaffolding strategy. This proof-of-concept study shows that LabLinking makes it possible to conduct collaborative HRI studies in remote laboratories and lays the first foundation for more in-depth research into robotic scaffolding strategies.
... FPs have been studied in different languages and from different perspectives. There are various explanations provided for filled pause production, their functions in speech, and their perception by listeners (Corley et al. 2007;Corley and Stewart 2008;Ferreira and Bailey 2004;Finlayson and Corley 2012;Fox Tree 2002;Lickley 2015;Navarretta 2015;O'Connell and Kowal 2005;Shriberg 2001;Clark and Fox Tree 2002;Stouten et al. 2006;Tottie 2016;Watanabe et al. 2008;etc.). Investigations have studied the incidence and characteristics of FPs during both L1 and L2 language acquisition, with typical speakers and also with atypical speakers (e.g., Bortfeld et al. 2001;Gayraud et al. 2011;Gósy et al. 2014;Hlavac 2011; characteristics of FPs during both L1 and L2 language acqui and also with atypical speakers (e.g., Bortfeld et al. 2001;Ga 2014;Hlavac 2011;de Jong and Bosker 2013;Merlo and Bar Research has demonstrated the frequent clustering of discour (Crible et al. 2017;Kosmala and Crible 2022). ...
... Each unit of speech from segments to intonation curves contains information. Listeners process both speech and non-speech information at the same time during speech perception and speech comprehension either consciously or unconsciously (Brodbeck et al. 2018;Finlayson and Corley 2012;Frank and Jaeger 2008;Gwilliams and Davis 2022;Jaeger 2010;etc.). Specific gaps like FPs occurring in the speech flow convey information about speech planning, speaker behavior, language properties, topic content, etc., as well as various messages for the listeners (e.g., Cossavella and Cevasco 2021;Fox Tree 2002;Fraundorf and Watson 2011;Kirjavainen et al. 2021;Levelt 1989;Local and Kelly 1986;O'Connell and Kowal 2005;Tottie 2016). ...
... Although FPs appear to interrupt the speech flow, they are in fact important components of speech production that are audible to the listener. Most studies define the main functions of FPs as providing extra time for the speaker to overcome speech planning or execution difficulties and providing pragmatic signals (e.g., turn-taking) for listeners and supplement various information contents (e.g., Arnold et al. 2003;Ferreira and Bailey 2004;Fox Tree 2002;Finlayson and Corley 2012;Levelt 1989;Roberts et al. 2009;Roggia 2012;Swerts 1998;Tottie 2014). ...
Filled pauses (i.e., gaps in speech production filled with non-lexical vocalizations) have been studied for more than sixty years in different languages. These studies utilize many different approaches to explore the origins, specific patterns, forms, incidents, positions, and functions of filled pauses. The present research examines the presence of filled pauses by considering the adjacent words and silent pauses that define their immediate positions as well as the influence of the immediate position on filled pause duration. The durations of 2450 filled pauses produced in 30 narratives were analyzed in terms of their incidence, immediate positions, neighboring silent pauses, and surrounding word types. The data obtained showed that filled pauses that were attached to a word on one side were the most frequent. Filled pauses occurring within a word and between two silent pauses were the longest of all. Hence, the durations of filled pauses were significantly influenced by the silent pauses occurring in their vicinity. The durations and occurrence of filled pauses did not differ when content or function words preceded the filled pause or followed it. These findings suggest that the incidence and duration of filled pauses as influenced by the neighboring words and silent pauses may be indicative of their information content, which is related to the processes of transforming ideas into grammatical structures.
... As an alternative conceptualization of disfluency features, fluencemes such as filled pauses "have the potential to serve both fluent and disfluent functions, depending on the configuration and context of use" (Kosmala & Crible, 2022, p. 9). In this way, fluencemes associated with breakdown fluency (e.g., silent pauses and filled gaps) may function as "task performing elements" (Kjellmer, 2003, p. 181), or discourse markers allowing speakers to buy time during lexical retrieval (Finlayson & Corley, 2012;Tavakoli & Wright, 2020), indicate shifts in topic (Kosmala & Crible, 2021) or turn-taking within conversation, or may be used as correction markers indicating that a reformulation will follow (Kjellmer, 2003). As a functional awareness of fluency develops within the L2, perhaps through heightened sociopragmatic knowledge, the increase in proficiency may benefit the language user in ways such as allowing them to better hold the floor during conversation (Du, 2013). ...
... Equally significant is the unique role of pause clusters in L2 learner speech and the insight that such hesitations provide into the underlying difficulty in L2 production (Kormos & Trebits, 2012;Sadri Mirdamadi & de Jong, 2015). From the beginning, the role of pause clusters in Alceste's oral data related to moments of overly taxed lexical retrieval (Finlayson & Corley, 2012) and the production of complex syntax (for a more detailed discussion, see Evans, 2019;Evans & Larsen-Freeman, 2020). As many complex syntactic structures, particularly those associated with subordination, extend the dependency distance between constituents, the overreliance on pause cluster sequences may equate to the increased demand of producing complex syntax (Mellow, 2008). ...
The present study investigated the development of second language fluency from the perspective of complex dynamic systems theory in an un-tutored adult learner of English as a second language. To complement traditional studies of second language (L2) utterance fluency, this study adopted a functional, usage-based perspective and sought to understand how fluency changes not just in terms of numerical increases or decreases in the frequency of disfluency features, but also how the discursive function of fluency changes over time. Data were collected from two oral tasks on a weekly basis for one academic year and were analyzed quantitatively and qualitatively in an attempt to achieve dynamic method integration (Hiver & Al-Hoorie, 2020). Though many disfluency features demonstrated either stagnation or even increases in frequency over the period of data collection, the way in which these features were leveraged by the learner changed over time. Specifically, pause clusters and self-repair acquired new roles associated with the emergence, and the subsequent monitor and repair, of novel syntactic structures. Results suggest that understanding L2 utterance fluency development is complex and requires more than simple frequency counts of disfluency features. Instead, definitions of L2 fluency must account for the discursive function of disfluencies and how these functions change as speakers develop proficiency in the L2.
... Pro dva základní typy hezitací v angličtině, pro hezitace vokalické a konsonantické, je v mnoha studiích užíván zápis uh a um; zároveň platí, že v těchto pracích je jazykem, v němž jsou hezitace identifikovány a analyzovány, právě angličtina (Johnson, Sabin 1986;Fox Tree 2001;Finlayson, Corley 2012 Hezitace v angličtině se dále zaznamenávají er (hezitace vokalická) a erm (hezitace konsonantická), a to typicky v příkladech z angličtiny britské (Corley 2007;Schachter et al. 1991 a další). Wieling et al. (2016) užívají záznam hezitací uh a um jako zastřešující označení pro dva základní typy hezitací (vokalickou a konsonantickou) s komentářem, že tento záznam je vhodný jak pro angličtinu americkou, tak pro angličtinu britskou. ...
The aim of this paper is to give an overview of the current knowledge about the phonetic realization and transcription of hesitations. Most authors distinguish two basic types of hesitations, namely vocalic and consonantal hesitations. Vocal hesitations are in the vast majority of cases (in different languages) realized as a neutral central vowel, schwa [ə], consonantal hesitations are realised as a consonant [m] without a final explosion. Some authors also mention the possible nasalization of vocalic hesitation [ə]. The two basic types of hesitations are most often transcribed as uh (vocalic hesitation) and um (consonantal hesitation). This method of transcription is mainly used by English writers who analyse (American) English. However, the repertoire of hesitation transcription varies from author to author, including with respect to the different languages that authors deal with when analyzing hesitations.
... According to Bühler, there are three ways in which linguistic signs function: they serve as symbols for concepts and objects from the outside world, as symptoms marking speaker state, and as signals indicating the speaker's intent. Since about the 1960s, fillers in particular have been looked at primarily as indications of verbal planning and self-monitoring on the part of the speaker (cf., e.g., Goldman-Eisler 1961;Shriberg 2001;Maclay and Osgood 1959;Finlayson and Corley 2012) and thus as examples of the symptom function of speech. Levelt (1989, p. 484) summarizes that "[t]he interjection 'er' apparently signals that at the moment when trouble is detected, the source of trouble is still actual or quite recent. ...
The forensic application of phonetics relies on individuality in speech. In the forensic domain, individual patterns of verbal and paraverbal behavior are of interest which are readily available, measurable, consistent, and robust to disguise and to telephone transmission. This contribution is written from the perspective of the forensic phonetic practitioner and seeks to establish a more comprehensive concept of disfluency than previous studies have. A taxonomy of possible variables forming part of what can be termed disfluency behavior is outlined. It includes the “classical” fillers, but extends well beyond these, covering, among others, additional types of fillers as well as prolongations, but also the way in which fillers are combined with pauses. In the empirical section, the materials collected for an earlier study are re-examined and subjected to two different statistical procedures in an attempt to approach the issue of individuality. Recordings consist of several minutes of spontaneous speech by eight speakers on three different occasions. Beyond the established set of hesitation markers, additional aspects of disfluency behavior which fulfill the criteria outlined above are included in the analysis. The proportion of various types of disfluency markers is determined. Both statistical approaches suggest that these speakers can be distinguished at a level far above chance using the disfluency data. At the same time, the results show that it is difficult to pin down a single measure which characterizes the disfluency behavior of an individual speaker. The forensic implications of these findings are discussed.
... A kitöltött szünetek jelenléte utalhat a fogalmi és a nyelvi megformálás háttérben zajló működésére (Beattie & Butterwoth, 1979;Levelt, 1989;Shriberg, 2001;Watanabe et al., 2008;Gósy et al., 2013;Gyarmathy, 2017), illetve megjelenhetnek más megakadásjelenségek, például a hibajelenségek és javításuk kísérlőjelenségeként (Gyarmathy, 2017) vagy a rejtett önmonitorozás felszíni jeleként (Christenfeld, 1996;Swerts, 1998;Clark & Fox Tree, 2002;Simpson, 2006;Finlayson & Corley, 2012;Gósy et al., 2013 elemek" aktivációja gátolt, és csupán az epizodikus elemeket strukturáló és ér-telmező "fogalmi keretekhez" (pl. "nem kedvel", "kiközösítettek") férnek hozzá (Conway, 2009). ...
Borderline personality disorder (BPD) is characterized by a pervasive pattern of instability of identity, emotions, and interpersonal relationships, and difficulty with emotional and impulse control. Due to the various co-occurrence patterns of symptoms and the interactions between them, the BPD population is relatively heterogeneous, making it difficult for clinicians to diagnose individuals. Since an individual's state of mind may to some extent be reflected in their speech behavior, an analysis of speech patterns may make a useful contribution to efforts of identifying the disorder. Our goal is to characterize spontaneous speech of BPD individuals based on the patterns of disfluencies and temporal parameters with the forced alignment method of automatic speech recognition, and to differentiate BPD individuals (N = 27) from healthy controls (N = 27) with a logistic regression statistical model. Our results have shown, that spontaneous speech of BPD individuals can be characterized by the frequency of silent pauses, filled pauses, and disturbances of grammatical encoding (grammatical errors and blendings), and the two groups can be differentiated by these features with 0.834 AUC.
... Following the work of Tottie (2011Tottie ( , 2014Tottie ( , 2016, the aim of the present research is to explore the pragmatic dimension of filled pauses across two different settings, by adopting a multimodal perspective. The following analyses focus on three specific functions served by filled pauses, which have been documented in the literature, mainly (1) planning and word searching (Tottie 2014(Tottie , 2016Jehoul, Brône & Feyaerts 2016;Finlayson & Corley 2012), (2) stalling and time-buying (Hieke 1981; and (3) turn-taking/holding (Beňuš 2009;Kjellmer 2003). The following analyses will show how these practices are not only performed vocally, but through the means of visible actions, such as gesture and gaze. ...
The present study aims to explore the status of filled pauses as pragmatic markers by taking into account their accompanying visual and gestural behavior. This aspect has not yet been widely explored, and the current study breaks new ground by demonstrating that the analysis of gaze and gesture can shed substantial light on the pragmatic functions of filled pauses and other pausing phenomena. Filled pauses (FPs) serve several pragmatic functions in speech, mainly planning but also turn-holding and emphasis, and their use is also highly determined by register and setting. This research explores the different pragmatic functions of FPs by analyzing their distribution in two different communication settings (conversation vs presentation setting), combining a quantitative and a qualitative methodology, following Kosmala & Crible’s (2021) study on the same data. Particular attention was paid to the co-occurring gestural activity of uh/ums and gaze behavior. Analyses show that the pragmatic functions of FPs are also embodied in kinetic activities which differ according to the setting: more pragmatic and referential ones were found during FPs in conversation than in the presentation setting, as well as more eye-contact, which reflects their potential communicative role during interactional sequences.
... Disfluency is an obstacle to speech that is not fluent or irregular. Various breaks, irregularities, or non-lexical vocables that occur in the flow of otherwise flu ent speech are referred to as speech disfluency, (Sukriana, Jufrizal, and Wahyuni, 2018;Finlayson & Corley, 2012). These include fillers such as grunts or non-lexical utterances (huh, uh, erm, um, well, so, and similar), false starts, which are words and sentences that are interrupted mid-utterance, phrases that are restarted or repeated, and repeated syllables, as well as repaired utterances, which are instances of speakers correcting their slips of the tongue or mispronunciations (before anyone else gets a chance to). ...
This study used a psycholinguistic approach and concentrated on speech disfluency in an Ellen Degeneres podcast. The goal of this study was to identify the most prevalent type of speech disfluency used in the Ellen Degeneres podcast by analyzing different types of speech disfluency. Because the data were gathered from documents, the research method was descriptive qualitative research with content analysis. The researchers chooses Podcast to air on July 3, 2021 with a duration of 31 minutes 45 seconds. The theory of speech disfluency by Clark and Clark (1977) was used to analyze the speech disfluency. The analysis revealed 77 speech inflections in Ellen Degeneres' Talk Show, including: Silent pauses (1%) and filled pauses (35%) as well as repetitions (10%), false starts (1%) and false starts (retraced) (0%), corrections (14%) and stutters (6%). Ellen Degeneres' filled pauses were the most prevalent type of speech disfluency in "Ellen Degeneres Explains Why She's Ending Her Show". Filled pauses used by speaker in "Ellen Degeneres Explains Why She’s Ending Her Show" to occurs because there is a pause to continue the words that will be spoken carefully. So, the conversation become run well. In conclusion, with this research we must occurs because there is a pause to continue the words that will be spoken carefully.
... The question of the linguistic status (are they similar to words?) of filler particles has been around for a long time in the literature (Clark and Fox Tree 2002;Keseling 1989), together with the question of whether they are rather symptoms or signals (Finlayson and Corley 2012;Walker et al. 2014), or both (Reitbrecht 2017), or both depending on their context (Kosmala and Crible 2022), being used either as fluent or disfluent constructions. At the same time, Kosmala and Crible (2022) argue against a word status of filled pauses. ...
The terms hesitation, planner, filler, and filled pause do not always refer to the same phonetic entities. This terminological conundrum is approached by investigating the observational, explanatory, and descriptive inadequacies of the terms in use. Concomitantly, the term filler particle is motivated and a definition is proposed that identifies its phonetic exponents and describes them within the linguistic category of particles. The definition of filler particles proposed here is grounded both theoretically and empirically and then applied to a corpus of spontaneous dialogues with 32 speakers of German, showing that in addition to the prototypical phonetic forms, there is a substantial amount of non-prototypical forms, i.e., 9.5%, comprising both glottal (e.g., [ʔ]) and vocal forms (e.g., [ɛɸ], [j̰ɛvə]). The grammatical classification and the results regarding the phonetic forms are discussed with respect to their theoretical relevance in filler particle research and corpus studies. The phonetic approach taken here further suggests a continuum of phonetic forms of filler particles, ranging from singleton segments to multi-syllabic entities.
In 7 experiments the authors investigated the locus of word frequency effects in speech production. Experiment 1 demonstrated a frequency effect in picture naming that was robust over repetitions. Experiments 2, 3, and 7 excluded contributions from object identification and initiation of articulation. Experiments 4 and 5 investigated whether the effect arises in accessing the syntactic word (lemma) by using a grammatical gender decision task. Although a frequency effect was found, it dissipated under repeated access to word's gender. Experiment 6 tested whether the robust frequency effect arises in accessing the phonological form (lexeme) by having Ss translate words that produced homophones. Low-frequent homophones behaved like high-frequent controls, inheriting the accessing speed of their high-frequent homophone twins. Because homophones share the lexeme, not the lemma, this suggests a lexeme-level origin of the robust effect.
Presents a standardized set of 260 pictures for use in experiments investigating differences and similarities in the processing of pictures and words. The pictures are black-and-white line drawings executed according to a set of rules that provide consistency of pictorial representation. They have been standardized on 4 variables of central relevance to memory and cognitive processing: name agreement, image agreement, familiarity, and visual complexity. The intercorrelations among the 4 measures were low, suggesting that they are indices of different attributes of the pictures. The concepts were selected to provide exemplars from several widely studied semantic categories. Sources of naming variance, and mean familiarity and complexity of the exemplars, differed significantly across the set of categories investigated. The potential significance of each of the normative variables to a number of semantic and episodic memory tasks is discussed. (34 ref) (PsycINFO Database Record (c) 2006 APA, all rights reserved).
Traditional accounts of language processing suggest that monologue – presenting and listening to speeches – should be more straightforward than dialogue – holding a conversation. This is clearly not the case. We argue that conversation is easy because of an interactive processing mechanism that leads to the alignment of linguistic representations between partners. Interactive alignment occurs via automatic alignment channels that are functionally similar to the automatic links between perception and behaviour (the so-called perception – behaviour expressway) proposed in recent accounts of social interaction. We conclude that humans are 'designed' for dialogue rather than monologue. Whereas many people find it difficult to present a speech or even listen to one, we are all very good at talking to each other. This might seem a rather obvious and banal observation, but from a cognitive point of view the apparent ease of conversation is paradoxical. The range and complexity of the information that is required in monologue (preparing and listening to speeches) is much less than is required in dialogue (holding a conversation). In this article we suggest that dialogue processing is easy because it takes advantage of a processing mechanism that we call 'interactive alignment'. We argue that interactive alignment is automatic and reflects the fact that humans are designed for dialogue rather than monologue. We show how research in social cognition points to other similar automatic alignment mechanisms.
Speakers often gesture in telephone conversations, even though they are not visible to their addressees. To test whether this effect is due to being in a dialogue, we separated visibility and dialogue with three conditions: face-to-face dialogue (10 dyads), telephone dialogue (10 dyads), and monologue to a tape recorder (10 individuals). For the rate of gesturing, both dialogue and visibility had significant, independent effects, with the telephone condition consistently higher than the tape recorder. Also, as predicted, visibility alone significantly affected how speakers gestured: face-to-face speakers were more likely to make life-size gestures, to put information in their gestures that was not in their words, to make verbal reference to their gestures, and to use more gestures referring to the interaction itself. We speculate that demonstration, as a modality, may underlie these findings and may be intimately tied to dialogue while being suppressed in monologue.
In conversation, two people inevitably know different amounts about the topic of discussion, yet to make their references understood, they need to draw on knowledge and beliefs that they share. An expert and a novice talking with each other, therefore, must assess each other's expertise and accommodate to their differences. They do this in part, it is proposed, by assessing, supplying, and acquiring expertise as they collaborate in completing their references. In a study of this accommodation, pairs of people who were or were not familiar with New York City were asked to work together to arrange pictures of New York City landmarks by talking about them. They were able to assess each other's level of expertise almost immediately and to adjust their choice of proper names, descriptions, and perspectives accordingly. In doing so, experts supplied, and novices acquired, specialized knowledge that made referring more efficient.
Statistical approaches to overdispersion, correlated errors, shrinkage estimation, and smoothing of regression relationships may be encompassed within the framework of the generalized linear mixed model (GLMM). Given an unobserved vector of random effects, observations are assumed to be conditionally independent with means that depend on the linear predictor through a specified link function and conditional variances that are specified by a variance function, known prior weights and a scale factor. The random effects are assumed to be normally distributed with mean zero and dispersion matrix depending on unknown variance components. For problems involving time series, spatial aggregation and smoothing, the dispersion may be specified in terms of a rank deficient inverse covariance matrix. Approximation of the marginal quasi-likelihood using Laplace's method leads eventually to estimating equations based on penalized quasilikelihood or PQL for the mean parameters and pseudo-likelihood for the variances. Im...
Statistical approaches to overdispersion, correlated errors, shrinkage estimation, and smoothing of regression relationships may be encompassed within the framework of the generalized linear mixed model (GLMM). Given an unobserved vector of random effects, observations are assumed to be conditionally independent with means that depend on the linear predictor through a specified link function and conditional variances that are specified by a variance function, known prior weights and a scale factor. The random effects are assumed to be normally distributed with mean zero and dispersion matrix depending on unknown variance components. For problems involving time series, spatial aggregation and smoothing, the dispersion may be specified in terms of a rank deficient inverse covariance matrix. Approximation of the marginal quasi-likelihood using Laplace's method leads eventually to estimating equations based on penalized quasilikelihood or PQL for the mean parameters and pseudo-likelihood for the variances. Implementation involves repeated calls to normal theory procedures for REML estimation in variance components problems. By means of informal mathematical arguments, simulations and a series of worked examples, we conclude that PQL is of practical value for approximate inference on parameters and realizations of random effects in the hierarchical model. The applications cover overdispersion in binomial proportions of seed germination; longitudinal analysis of attack rates in epilepsy patients; smoothing of birth cohort effects in an age-cohort model of breast cancer incidence; evaluation of curvature of birth cohort effects in a case-control study of childhood cancer and obstetric radiation; spatial aggregation of lip cancer rates in Scottish counties; and the success of salamander matings in a complicated experiment involving crossing of male and female effects. PQL tends to underestimate somewhat the variance components and (in absolute value) fixed effects when applied to clustered binary data, but the situation improves rapidly for binomial observations having denominators greater than one.
In spontaneous speaking, the is normally pronounced as thuh, with the reduced vowel schwa (rhyming with the first syllable of about). But it is sometimes pronounced as thiy, with a nonreduced vowel (rhyming with see). In a large corpus of spontaneous English conversation, speakers were found to use thiy to signal an immediate suspension of speech to deal with a problem in production. Fully 81% of the instances of thiy in the corpus were followed by a suspension of speech, whereas only 7% of a matched sample of thuhs were followed by such suspensions. The problems people dealt with after thiy were at many levels of production, including articulation, word retrieval, and choice of message, but most were in the following nominal. © 1997 Elsevier Science B.V. All rights reserved.
The delay markers (DMs) uh and um are often used by adult English speakers to indicate that an upcoming pause is due to a speech disruption, not the end of a conversational turn. Moreover, uh and um indicate different degrees of disruption (Clark & Fox Tree, 2002). Thus, it appears that children must learn how to use DMs appropriately. In the current study we examined DM use in elicited speech samples from 24 3- and 4-year-old children. We found that pauses following DMs were longer than those not following a DM, but that there was no difference between the pauses following uh and um. Children at this age, then, appear to understand the basic use of DMs, but do not yet differentiate between them.