Disfluency Types Research Articles

The effect of disfluency on memory for what was said.

Disfluencies such as pauses, "um"s, and "uh"s are common interruptions in the speech stream. Previous work probing memory for disfluent speech shows memory benefits for disfluent compared to fluent materials. Complementary evidence from studies of language production and comprehension have been argued to show that different disfluency types appear in distinct contexts and, as a result, serve as a meaningful cue. If the disfluency-memory boost is a result of sensitivity to these form-meaning mappings, forms of disfluency that cue new upcoming information (fillers and pauses) may produce a stronger memory boost compared to forms that reflect speaker difficulty (repetitions). If the disfluency-memory boost is simply due to the attentional-orienting properties of a disruption to fluent speech, different disfluency forms may produce similar memory benefit. Experiments 1 and 2 compared the relative mnemonic benefit of three types of disfluent interruptions. Experiments 3 and 4 examined the scope of the disfluency-memory boost to probe its cognitive underpinnings. Across the four experiments, we observed a disfluency-memory boost for three types of disfluency that were tested. This boost was local and position dependent, only manifesting when the disfluency immediately preceded a critical memory probe word at the end of the sentence. Our findings reveal a short-lived disfluency-memory boost that manifests at the end of the sentence but is evoked by multiple types of disfluent forms, consistent with the idea that disfluencies bring attentional focus to immediately upcoming material. The downstream consequence of this localized memory benefit is better understanding and encoding of the speaker's message. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Decontextualized Utterances Contain More Typical and Stuttering-Like Disfluencies in Preschoolers Who Do and Do Not Stutter.

Stuttering-like disfluencies (SLDs) and typical disfluencies (TDs) are both more likely to occur as utterance length increases. However, longer and shorter utterances differ by more than the number of morphemes: They may also serve different communicative functions or describe different ideas. Decontextualized language, or language that describes events and concepts outside of the "here and now," is associated with longer utterances. Prior work has shown that language samples taken in decontextualized contexts contain more disfluencies, but averaging across an entire language sample creates a confound between utterance length and decontextualization as contributors to stuttering. We coded individual utterances from naturalistic play samples to test the hypothesis that decontextualized language leads to increased disfluencies above and beyond the effects of utterance length. We used archival transcripts of language samples from 15 preschool children who stutter (CWS) and 15 age- and sex-matched children who do not stutter (CWNS). Utterances were coded as either contextualized or decontextualized, and we used mixed-effects logistic regression to investigate the impact of utterance length and decontextualization on SLDs and TDs. CWS were more likely to stutter when producing decontextualized utterances, even when controlling for utterance length. An interaction between decontextualization and utterance length indicated that the effect of decontextualization was greatest for shorter utterances. TDs increased in decontextualized utterances when controlling for utterance length for both CWS and CWNS. The effect of decontextualization on TDs did not differ statistically between the two groups. The increased working memory demands associated with decontextualized language contribute to increased language planning effort. This leads to increased TD in CWS and CWNS. Under a multifactorial dynamic model of stuttering, the increased language demands may also contribute to increased stuttering in CWS due to instabilities in their speech motor systems.

Bilingualism as a risk factor for false reports of stuttering in the Early Childhood Longitudinal Study (ECLS-K:2011).

Bilingualism has historically been claimed to be a risk factor for developmental stuttering. The Early Childhood Longitudinal Study, Kindergarten Class of 2010-11 (ECLS-K:2011) ostensibly contains evidence to test that claim. We analyze data from monolingual and bilingual children in Kindergarten through fifth grade in the ECLS-K:2011. The prevalence, male/female ratio, and onset and recovery of reported stuttering in the ECLS are inconsistent with widely-accepted clinical reports of stuttering. We argue that the reported figures may be misleading. We discuss some factors that may inflate the reported prevalence, including a lack of awareness of the difference between stuttering vs. normal disfluencies, and the informal usage of the word "stuttering" on the part of teachers and parents to describe typical disfluencies.

Effects of task variations on language productivity, syntactic complexity, and stuttering in children who stutter

ABSTRACT Purpose. The aim of this preliminary study was to compare the effects of variations in task demands on: (a) language productivity in sentences categorised into stuttered vs. non-stuttered sentences; (b) syntactic complexity in stuttered vs. non-stuttered sentence categories, and (c) stuttering and typical disfluencies in school-age children who stutter (CWS). Language Sample Analysis (LSA) was conducted on samples from three tasks – Conversation, fable retell, and critical thinking based on the fables. Methods. Participants were 14 CWS categorised into younger (9 to 12-year-olds, n = 8) and older age groups (13 to 15-year-olds, n = 6). The Computerized Language ANalysis program was used to conduct language and disfluency analyses. Repeated measures analysis of variance and nonparametric statistical analyses were used to investigate: (a) Language productivity in total number of words and sentences by task and sentence category; (b) Syntactic complexity at the word- (use of metacognitive verbs), phrase- (use of verb phrases), and utterance (Mean Length of Utterances in words, MLUw) levels by task and sentence category; and (c) Disfluencies measured using % stuttered syllables (%SS) and % typical disfluencies (%TD). Results and Conclusions. Task effects in language productivity did not differ by sentence category and suggested limited influences of propositionality and volubility in stuttering. In contrast, higher syntactic complexity was obtained in the stuttered compared to non-stuttered sentences at the word, phrase, and utterance levels and it was the same task – conversation, that elicited the effect. Additionally, variations in task demands did not result in significant differences in %SS. The findings inform assessment planning with the selection of tasks guided by task demands and assessment requirements.

Pausing and fluency in speech of patients with relapsing-remitting multiple sclerosis

ABSTRACT Multiple Sclerosis (MS) causes a variety of symptoms in speech production, such as more frequent pauses and an increase in the duration of pauses in the speech. However, there is almost no data on whether the disease affects speech fluency in other ways, such as changes in the frequency of disfluencies in speech. The main question of this study is the following: if we examine speech fluency in speech tasks requiring different cognitive load, will there be a difference between patients and controls? Twenty people with relapsing-remitting MS (3 men and 17 women) and 20 age- and education-matched control speakers (4 men and 16 women) participated in the study. Speech samples were recorded with each participant in three speech tasks: 1) spontaneous narratives about their own lives, 2) narratives about their previous day, and 3) narrative recalls based on a heard text. In the speech samples, pauses and disfluencies were annotated and the duration of pauses was measured. Then, the frequency of pauses and disfluencies were calculated and the types of disfluencies were examined. The results show that there are differences in the frequency and duration of pauses between people with MS and controls. However, there were no significant differences in the frequency of disfluencies between the groups. The same types of disfluencies occurred in the same frequency in both groups. The results help to better understand the speech production processes in MS.

A cross-sectional investigation of disfluencies in typically developing Spanish-English bilingual children

PurposeThis study examined the language skills and the type and frequency of disfluencies in the spoken narrative production of typically developing Spanish-English bilingual children. MethodA cross-sectional sample of 106 bilingual children (50 boys; 56 girls) enrolled in kindergarten through Grade 4, produced a total of 212 narrative retell language samples in English and Spanish. A specialized fluency coding system was implemented to index the percentage of total (%TD) and stuttering-like disfluencies (%SLD) in each language. Large-scale reference databases were used to classify children’s dual language proficiency profiles (balanced, English dominant, Spanish dominant) based on language sample analysis measures of morphosyntax and lexical diversity. ResultsThe bilingual Spanish-English children in this study did not demonstrate significant cross-linguistic differences for mean %TD or %SLD. However, the mean %TD and %SLD in both languages exceeded the risk threshold based on monolingual English-speaking norms. English dominant bilingual children demonstrated significantly lower %TD in English than Spanish. Spanish dominant children demonstrated significantly lower %SLD in Spanish than English. ConclusionsThis study included the largest sample size of bilingual Spanish-English children investigated to date from a fluency perspective. The frequency of disfluencies was found to be variable across participants and change dynamically as a function of grade and dual language proficiency profiles, indicating the need for studies that employ larger sample sizes and longitudinal designs.

Comparative Analysis of the Temporal Aspects of Word-Final Disfluencies, Stuttering-Like Disfluencies, and Nonstuttering-Like Disfluencies

Purpose: Repetitions at the end of words, known as word-final disfluencies (WFDs), have long been documented in the speech of individuals with and without other communication disorders. Although these disfluencies are often compared to repetitions at the beginning of words observed in stuttering, it has been suggested that WFDs differ from stuttering in their salient features. WFDs have been observed to present more as nonstuttering-like disfluencies (NSLDs; interjections, phrase repetitions, revisions) than as stuttering-like disfluencies (SLDs) in terms of tension and the individual's awareness of their disfluencies. An acoustic comparison of the different types of disfluency has not yet been explored. The purpose of this study was to compare the durational characteristics of repetitions in individuals with WFDs to those in individuals who exhibit SLDs and NSLDs. Method: Speech samples from three school-age boys who presented with WFDs, three school-age boys who stutter, and three preschool-age boys who presented with NSLDs often seen in normal language development were recorded and analyzed. The silent duration between the parent word and the repetition unit of each participant's disfluencies were measured using acoustic analysis software and compared across groups. Results: Results indicated that the silent duration between the word and the repeated segment of the disfluencies produced by each group was significantly different from each other. Conclusion: Descriptively, the silent duration of WFDs was closer to that of NSLDs than to that of SLDs, suggesting that WFDs may be more closely related to the disfluencies of developing language than to those of stuttering.

Semantic interference affects speech production by increasing disfluencies, not errors

Several studies have shown that different types of disfluency occur depending on the language production stage at which people experience difficulties. The current study combined a network task and a picture–word interference task to analyse whether lexical-semantic difficulty triggers errors and disfluencies in connected-speech production. The participants produced more disfluencies in the presence of a semantically related distractor word than an unrelated distractor word, while few semantic errors were made. These results support the hypothesis that difficulties at distinct stages of language production lead to distinct patterns of disfluency, with lexical-semantic difficulties leading to self-corrections and silent pauses. The results also have implications for the role of the monitoring system in connected-speech production.

Stalling for Time: Stall, Revision, and Stuttering-Like Disfluencies Reflect Language Factors in the Speech of Young Children.

Disfluencies can be classified as stuttering-like disfluencies (SLDs) or typical disfluencies (TDs). Dividing TDs further, stalls (fillers and repetitions) are thought to be prospective, occurring due to planning glitches, and revisions (word and phrase revisions, word fragments) are thought to be retrospective, occurring when a speaker corrects language produced in error. In the first study assessing stalls, revisions, and SLDs in matched groups of children who stutter (CWS) and children who do not stutter (CWNS), we hypothesized that SLDs and stalls would increase with utterance length and grammaticality but not with a child's expressive language level. We expected revisions to be associated with a child having more advanced language but not with utterance length or grammaticality. We hypothesized that SLDs and stalls (thought to be planning-related) would tend to precede grammatical errors. We analyzed 15,782 utterances from 32 preschool-age CWS and 32 matched CWNS to assess these predictions. Stalls and revisions increased in ungrammatical and longer utterances and with the child's language level. SLDs increased in ungrammatical and longer utterances, but not with overall language level. SLDs and stalls tended to occur before grammatical errors. Results suggest that both stalls and revisions are more likely to occur in utterances that are harder to plan (those that are ungrammatical and/or longer) and that, as children's language develops, so do the skills they need to produce both stalls and revisions. We discuss clinical implications of the finding that ungrammatical utterances are more likely to be stuttered.

Comparative analysis of various feature extraction techniques for classification of speech disfluencies

Speech plays a vital role in communication, from expressing oneself, to utilizing speech-based platforms, speech is a necessity. Any disruption in speech is referred to as disfluency, and can impact one's quality of life. This paper presents an experimental study on various techniques for the detection and classification of speech disfluencies. Six different types of disfluencies are examined in this paper, namely Interjection, Sound Repetition, Word Repetition, Phrase Repetition, Revision and Prolongation (6 classes). However, this paper also goes a step further by including the clean speech signals as an added class alongside the six disfluencies, thereby making this work more robust with 7 classes. Various machine learning approaches have been investigated on the University College London Archive of Stuttered Speech (UCLASS) dataset; a standard disfluency dataset generated by University College London (UCL). Five different feature extraction techniques viz. Mel Frequency Cepstral Coefficients (MFCC), Linear Predictive Cepstral Coefficients (LPCC), Gammatone Frequency Cepstral Coefficients (GFCC), Mel-filterbank energy features, and Spectrograms have been used. Comparative analysis of various classifiers shows that MFCC, GFCC, and Spectrograms achieved greater than 90% accuracy on both 6 and 7 classes with the kNN classifier. As a future scope to this study, the authors aim to focus on tackling the challenges of detecting multiple disfluencies present simultaneously in a speech sample.

Exploratory Examination of Speech Disfluencies in Spoken Narrative Samples of School-Age Bidialectal Children

This study examined the relationship between school-age children's speech disfluencies and the use of and variation of Mainstream American English (MAE) and African American English (AAE). Given that bilingual children may present with notably more speech disfluencies than monolingual children, it was hypothesized that bidialectal speaking children (i.e., those that use both MAE and AAE) may exhibit higher speech disfluencies, as compared to children who speak mainly MAE and those who mainly speak AAE. It was also hypothesized that bidialectal speaking children would exhibit a greater variety of speech disfluency types when compared to the other two dialect groups (i.e., MAE and AAE). School-age children (n = 61) with typical development and fluency were classified into three dialect groups: MAE speakers (n = 21), bidialectal MAE-AAE speakers (n = 11), and AAE speakers (n = 29). Tell-retell narrative samples were elicited from each participant using a wordless picture book. Speech disfluencies exhibited during these narrative samples were examined for frequency of stuttering-like and nonstuttering-like speech disfluencies and type of speech disfluency. Findings indicated that bidialectal speaking children do not present with a higher frequency of speech disfluencies when compared to children who speak MAE and children who speak AAE. Additionally, there were no differences in the types of speech disfluencies exhibited by the different dialect groups. Unexpected findings of this study nullify both hypotheses and suggest that bidialectalism, in comparison to bilingualism, has less of an impact on speech fluency. Findings provide evidence that bidialectal speaking children are not at an increased risk for a misdiagnosis of stuttering. Clinically, these preliminary findings provide some scientific validity and specification to the appropriateness of using already established diagnostic criteria commonly used for stuttering with dialect speakers.

Identifying the Speech Errors in a Talkshow From a Podcast: A Case on Speech Disfluency

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.

Usage of the Speech Disfluency Detection Method for the Machine Translation of the Transcriptions of Spoken Language

Neural machine translation falls into the category of natural language processing tasks. Despite the availability of a big number of research papers that are devoted to the improvement of the quality of the machine translation of documents, the problem of the translation of the spoken language that contains the elements of the disfluency speech is still an actual task, especially for low-resource languages like the Ukrainian language. In this paper, the problem of the neural machine translation of the transcription results of the spoken language that incorporate different elements of the disfluency speech has been considered in the case of the translation from the English language to the Ukrainian language. Different methods and software libraries for the detection of the elements of disfluency speech in English texts have been analyzed. Due to the lack of open-access corpora of the speech disfluency samples, a new synthetic labeled corpus has been created. The created corpus contains both the original version of a document and its modified version according to the different types of speech disfluency: filler words (uh, ah, etc.) and phrases (you know, I mean), reparandum-repair pairs (cases when a speaker corrects himself during the speech). The experimental verification of the effectiveness of the usage of the method of disfluency speech detection for the improvement of the machine translation of the spoken language has been performed for the pair of English and Ukrainian languages. It has been shown that the current state-of-the-art neural translation models cannot produce the appropriate translation of the elements of speech disfluency, especially, in the reparandum-repair cases. The results obtained may indicate that the mentioned method of disfluency speech detection can be used for the previous processing of the transcriptions of spoken dialogues for the creation of coherent translations by the usage of the different models of neural machine translation.

The Speaking Disfluencies in the University Students’ Classroom Conversation

The main topic as the research focus in this study deals with the investigation of the category of the speaking disfluencies of the students of the English Study Program. To explore the types of the speaking disfluency produced by the students of the English Study Program of FKIP Universitas HKBP Nommensen Pematangsiantar in their classroom conversationThe design of this research is categorized to a qualitative design. The technique of the data collection is by participant observation. By this technique the researcher functions as the key instrument in the data collection. In collecting the data the researcher is acting as a substitute lecturer giving a subject in Interpersonal Speaking The results of the study present an analysis of data on the types of speech impediments produced by students of the English Study Program at the University of HKBP Nommensen Pematangsiantar FKIP in their speeches in speaking class. And find analysis data on how each type of speech disfluency is used by students of the English Language Study Program at the HKBP Nommensen University FKIPThe students of English program FKIP Nommensen University use all of the type’s disfluency such as Filler or Repairs, Repetition, Hesitation and False Starts. They use Fillers or Repair while they are having a speech in front of their classroom and occurs every five to eight words.

Linguistic disfluencies in Russian-speaking typically and atypically developing children: individual variability in different contexts

ABSTRACT Disfluency in children and adults seems to occur like errors of speech but, at the same time, is an essential feature of spontaneous (unprepared) speech. The present study aimed to evaluate linguistic disfluencies in typically and atypically developing Russian-speaking children from the perspective of the dynamic adaptive model of self-monitoring in speech production. The study collected four language samples from 10 six-year-old children with developmental language disorder and 14 typically developing peers: two storytelling tasks, structured conversation, and a play argument. After transcribing audio-recordings and marking linguistic disfluencies, the authors conducted structured distributional analysis. The distribution of several indexes of disfluency was estimated to assess the prevalence and profiles of different (sub)types of disfluencies. The disfluency rate statistics were similar between the typically developing children and children with developmental language disorder. The distributional indexes score showed that tasks significantly impacted the rate of different (sub)types of disfluencies. Task-related patterns in a set of the distributional indexes significantly distinguished the groups. Thus, changes in the disfluency profile related to different external factors, as a sign of a flexibility of an adaptive self-monitoring system, may be limited in children with developmental language disorder.

A Comprehensive Analysis of Speech Disfluencies in Autistic Young Adults and Control Young Adults: Group Differences in Typical, Stuttering-Like, and Atypical Disfluencies.

The purpose of this study was to examine the nature of speech disfluencies in autistic young adults and controls by using a wide-range disfluency classification of typical disfluencies (TD; i.e., filled pauses, revisions, abandoned utterances, and multisyllable word and phrase repetitions), stuttering-like disfluencies (SLD; i.e., sound and syllable repetitions, monosyllable word repetitions, prolongations, blocks, and broken words), and atypical disfluencies (AD; i.e., word-final prolongations and repetitions and atypical insertions). Thirty-two autistic young adults and 35 controls completed a narrative telling task based on socially complex events. Frequencies of total disfluencies, TD, SLD, AD and stuttering severity were compared between groups. The overall frequency of disfluencies was significantly higher in the autistic group and significant between-group differences were found for all disfluency categories. The autistic group produced significantly more revisions, filled pauses, and abandoned utterances, and each subtype of SLD and AD than the control group. In total, approximately every fourth autistic participants scored at least a very mild severity of stuttering, and every fifth produced more than three SLD per 100 syllables. Disfluent speech can be challenging for effective communication. This study revealed that the speech of autistic young adults was highly more disfluent than that of the controls. The findings provide information on speech disfluency characteristics in autistic young adults and highlight the importance of evaluating speech disfluency with a wide-range disfluency classification in autistic persons in order to understand their role in overall communication. The results of this study offer tools for SLPs to evaluate and understand the nature of disfluencies in autistic persons.

Exploring the impact of directionality on disfluencies in simultaneous interpreting

This paper focusses on the concept of directionality in interpreting. The main aim of this study is to investigate four types of disfluencies in simultaneous interpreting performed by students and the impact of pausal phenomenon on directionality. Disfluencies being inseparable element of speech are the factor that has a substantial influence on interpreting process including directionality. The participants in this study were 12 advanced interpreting students who interpreted simultaneously two texts from A language into B language and two texts from B language into A language. Their outputs were later analysed taking into consideration four types of disfluencies and verified how they influence directionality. This study offers some insights into occurrences of disfluencies. Research findings show that students display tendency to produce more disfluencies while interpreting into A language. In addition, the obtained results highlight the necessity to focus not only on fluency but also on disfluencies while interpreting training to improve quality in simultaneous interpreting in both directions.

DISFLUENCIES AND THEIR GESTURE PROFILES: THE ANALYSIS OF SIMULTANEOUS INTERPRETING FROM L1 TO L2

The article elaborates on the specifics of simultaneous interpreting as a highly demanding cognitive task which is reflected in the presence of speech disfluencies and co-occurring gestures. The analysis of simultaneous interpreting showed the differences in the distribution of speech disfluencies and demonstrated some correlations between these disfluencies and gestures. The data was obtained from 24 simultaneous interpreters, 12 per each group (Russian-English and Russian-German). The corpus was annotated in ELAN program and analyzed using quantitative and statistical methods. The results suggest that in simultaneous interpreting there are certain differences in the distribution of disfluencies, and in the ways they are accompanied by gestures. Such differences are related to the target language, namely, whether it is English or German. Thus, Russian-English interpreting demonstrated a higher variety of disfluencies accompanied by gestures in comparison to Russian-German interpreting as was shown in the results of the qualitative analysis. For adapters and pragmatic gestures significant statistical co-dependences were found between types of disfluencies and gestures both in Russian-English and Russian-German corpora, however these correlations played out differently for the two target languages. Almost no significant differences were found for the representational and deictic gestures: they did not frequently co-occur with the speech disfluencies in our data base - probably, due to their higher conceptual purport and to the increase of cognitive load during this type of activity.

A Comparison of Speech Dysfluencies Before and After Sports-Related Concussion

<h3>Research Objectives</h3> The purpose of this study is to compare the number and type of speech disfluencies that occur in college-aged athletes at baseline compared to after a sports-related concussion. <h3>Design</h3> Pre-post (repeated measures). <h3>Setting</h3> University. <h3>Participants</h3> A total of 230 Division I student athletes participated in pre-season speech testing. Twelve students (18-22 years) who sustained a concussion participated in speech testing in the days following concussion. Picture description speech samples were independently coded by three trained raters for 17 error types in the traditional categories of errors defined in fluency analysis. <h3>Interventions</h3> Not Applicable. <h3>Main Outcome Measures</h3> Error types were sorted into three categories: stuttering-like dysfluencies (part-word repetition, single-syllable whole-word repetition, prolongation, block), articulation errors (substitution, distortion, addition, omission), and other dysfluencies (multisyllable whole-word repetition, revision, abandoned or incomplete segment, interjection, filler, phrase repetition, pauses). <h3>Results</h3> Within-subjects analysis comparing the difference in percent error scores at baseline and post-concussion revealed significant differences for interjections (t(11)=-2.678, p< .05). The Other Disfluency category was also significantly different (t(11)= -2.735, p< .05), with more errors occurring after a concussion. No change in the Stuttering-Like Disfluency (t(11)= -0.799, p>.05) or Articulation Error category (t(11)=-0.045, p>.05) was found. <h3>Conclusions</h3> As predicted, pre-post comparisons demonstrated an overall increase in total number of speech errors after a concussion (Mean Baseline = 11.5 errors; Mean Concussion = 15.9 errors). Results demonstrate that specific speech fluency changes occur following mild sports-related concussions. Future studies should consider speech as a diagnostic tool for concussion. <h3>Author(s) Disclosures</h3> Funding for this work is through an internal grant by Seton Hall University. No other relevant disclosures.

Speech Disfluencies of Non-Native Speakers of English in TED Talk Scripts

To speak fluently in the target language, English, has remained a challenge for learners, particularly those who are non-native speakers of English. When speaking, therefore, learners should be aware of speech disfluencies. This paper, accordingly, investigates the speech disfluencies of non-native speakers in their presentations. Data were collected from TED Talk scripts and were then analysed and discussed, based on categories of disfluency. The findings of this document analysis reveal five types of disfluency in the data, namely unfilled pause, filled pause, repetition, deletion, and substitution. It is expected that learners of English at university level can benefit from the study results concerning speech disfluencies to improve their speaking proficiency. Keywords: Disfluency; filled pause; non-native speaker; TED talk