2003 IFA Congress: Montreal, Canada

Aspects of Normally Fluent Speech in Brazilian Adults

Claudia Regina Furquim de Andrade, Femanda Chiarion Sassi,  Daniela Veronica Zackiewicz
Department of Physiotherapy, Speech-Language and Hearing Science and Occupational Therapy, School of Medicine, University'ofSc'io Paulo, Rua Cipotanea 51, Cidade Universitaria, Siio Paulo - S.P., 05360-160, Brazil


This study aimed to describe the normally fluent speech of Brazilian Portuguese speaking adults who do not stutter. Speech samples of 30 adults, 13 males and 17 females, whose ages ranged from 20 to 43 years, and had no history of any communication disorders were analyzed and the following measures obtained: speech typology of disruptions (typical and less typical disruptions), speech rate (syllables and words per minute), and rate of disruptions (percentage of speech discontinuity and percentage of stuttered syllables). Mean values and confidence levels and the clinical implications of these findings are presented.

  1. Introduction
Fluency can be defined as a continuum of smooth speech flow (Starkweather & Givens- Ackerman, 1997), which is a product of a harmonic integration of hearing, language and speech processes, as well as those in control centers of emotions and memory (Andrade, 2000). Fluency varies not only between different individuals but also within the same person in different situations (Van Riper & Emerick, 1990). This variability may be linked to prior knowledge of the topic or theme, emotions involved in speaking, and type of speech situation (Goldman-Eisler, 1961; Pais, 1986; Starkweather & Givens-Ackerman, 1997; Zackiewicz, 1999).

Disruptions in speech flow can be analyzed using different measures. Campbell and Hill (1994) described typical disfluencies in normally fluent speech as: interjections, hesitations, revisions, incomplete words, and phrase and word repetitions. Such events can result from linguistic uncertainties and imprecision or may be used to improve the comprehension of the message (Andrade, 1999; Bloodstein, 1979; Perkins, 1983; Starkweather & Givens-Ackerman, 1997). In contrast, disruptions such as syllable and sound repetitions, prolongation, blocks and long pauses which are usually associated with stuttering and other fluency disorders, occur infrequently in normally fluent speech (Andrade, 2000; Andrews et al., 1983; Barbosa & Chiari, 1998; Campbell & Hill, 1994; Johnson, 1961; Kelly, 1994; Van Riper & Emerick, 1990; Wingate, 1964).

The frequency at which normal disfluencies (more typical) or stuttered events (less typical) occur in speech is considered to be an important measure in fluency analysis (Johnson, 1961). Less typical events, or 2% of stuttered syllables, is often considered to be the maximum for normal speech (Bloodstein, 1979; Franken et al., 1995; Howell, 1997; Ingham & Riley, 1998; Kelly, 1994; Perello, 1995; Ryan, 1992; Schwartz & Conture, 1988; Throneburg et al., 1994; Yairi & Ambrose, 1992;

Yairi et al., 1993), whereas 4% stuttered syllables indicates that stuttering or some other disfluency disorder may be present (Yairi, 1997). Speech continuity can be understood as the correspondence between a speaker’s intention and how it is produced. It can be directly assessed as total time spent with non-intended elements divided by the total duration of what was said, which includes both non-intended and intended elements. The result is the percentage of speech time discontinuity (%STD) which is a measure of speech fluency that includes both the frequency and duration of discontinuities. As suggested by the literature (Starkweather & Givens-Ackerman, 1997), when a normally fluent speaker speaks, listeners infer the intended message although non-intentional elements are also present. Thus, listeners usually pay little attention to aspects of speech that are not intentional, and non intentional elements are brought to listener’s attention only if they occur excessively or are in some way unusual. For example, they will be unnoticed if the frequency of repetitions do not reach 15% of the message and if the interjections do not reach 20%.

The literature also advocates measuring speech rate at two levels, the number of words per minute (WPM) or number of syllables per minute (SPM). The word level indexes the rate at which a person is capable of producing information. For Lutz & Mallard (1986), 167.7 WPM or 220.6 SPM was the average for male speakers during conversation. The averages female speakers is lower (151.2 WPM and 204.2 SPM) and their conversation is also marked by the use of longer words. The syllable level refers to articulatory rate, which is the rate at which a person can move speech production structures. Although this measure is not free from the influence of linguistic planning and sentence formulation, articulatory rate measures the person’s motor control of speech. Articulatory rate depends on two variables, the rate at which syllable are initiated and their coarticulation. As speech rate increases, so does articulatory overlapping which can reduce intelligibility at high rates of speech. The authors’ normal production approximates 5 to 6 syllables per second and durations, around 20 milliseconds per syllable.

The aim of the present study was to describe the normally fluent speech of Brazilian Portuguese speaking adults who do not stutter. The measures obtained will be important not only to future research of normal speech, but also to clinical practice with individuals who stutter.

2 Method

Participants in this study were volunteers who regarded themselves as being fluent. Thirty subjects were assessed, 13 males and 17 females, whose ages ranged from 20 to 43 years from varying ethnic and racial backgrounds; and variable levels of literacy. Only 6.7% of the subjects finished junior school, 26.7% finished high school, 53.3% were attending or had finished college, and 13.3% had post- graduate education.

Criteria for inclusion in the study were: (a) no history of neurologic, hearing, cognitive or visual deficits or any communication disorders; (b) evidence of at least 75% adequacy in the applied screening; (c) speech samples within normal limits according to the Systematic Disfluency Analysis (Campbell & Hill, 1994). The study had prior approval from the Research Ethics Committee of the Institution (CAPPESC HCFMUSP 34997) and informed consent was obtained from all patients.

The following equipment was used to gather and analyze data: digital videocamera (Gradiente, GC-16C); tripod, microphone (Leson, ML8), videotapes (VHS-C, J VC, TC-30), digital chronometer (Casio, CPW-200), television , videocassette recorder and computer (Toshiba Satellite Pro 425 CDT). A picture stimulus was used to elicit monologues.

Subjects signed an informed agreement protocol after receiving information about the research projects, and then a case history was obtained, followed by a speech screening and a videotaped sample of speech during a monologue elicited by a picture stimulus. This method for collecting speech samples was selected because of its use during clinical assessments of stuttering and other disorders, when first analyzing cases as well as follow-up assessments of treatment. The use of a visual stimulus to elicit speech monologues allowed data to be collected without the influence of a conversation partner. 378 Theory, research and therapy in fluency disorders

Analyses of speech samples followed the methodology described by Andrade (1999, 2000) and Zackiewicz (1999) to obtain measures of typical and less typical disfiuencies, WPM and SPM, and percentages of speech discontinuity and stuttered syllables.

Judges for this study were three researchers with academic and professional experience in stuttering, in the issues surrounding stuttering measurements, and in treatment and research.

Speech samples were analyzed independently by each judge and results were compared at a later time. Interjudge reliability was obtained for SPM and WPM, and for identifying typical and less typical disfluencies. Interjudge reliability ranged from 0.91 to 0.96 (Sander, 1961).
  1. Results
Each speech sample was analyzed in terms of the six measures just listed. The mean values for each measure were used to draw inferences about the mean values of the population. The Confidence Level Test (CL), which determines the interval within which the mean value of the population lies at a confidence of 95% was used to determine these intervals (Rosner, 1986).


Table 1. Mean values and Confidence Intervals of typical and less typical disfluent events for females, males, and the total sample

Table 1 displays the mean numbers of typical and less typical speech disfluencies. The average number of typical disfluent events in 200 syllable samples was 16.3, and the population’s mean value is estimated to lie between 11.8 and 20.6. In contrast, the average number of less typical disfluent events was 0.9, with a confidence interval of 0.1 to 2.0. These means were higher for males than females as were the confidence intervals for both typical and less typical disfluencies. These findings suggest that the sex of individuals should be considered in fluency assessments in draw up inferences about typical and less typical disfluent events.


Table 2. Mean percentages and Confidence Intervals Levels of percentages of speech discontinuity (%SD) and stuttered syllables (%SS)

The mean percentages and confidence intervals for speech discontinuity and syllables stuttered are shown in Table 2. Mean speech discontinuity for the sample was 8.62% with a confidence interval for the population of 7.3% to 9.9%. This suggests that speech can include both typical and less typical disruptions and still be classified as normal. However, if the percentage of stuttered syllables is much higher than 0.7%, speech may not be perceived as normal.

Small differences between females and males are again present with those for speech discontinuity (males higher) being the largest.


Table 3. Mean value and Confidence Intervals of syllables minute (SPM) and words per minute (WPM)

Speech rate measures are presented in Table 3. Mean SPM was 237.67 with a confidence interval for the population of 218.8 to 256.5. The mean index for information production was 128.79 WPM with a confidence interval of 117.3 to 140.3 WPM. Females in this sample evidenced articulatory and information production rates higher than the males.

  1. Discussion
Although most studies of speech fluency focus on disordered speech, especially stuttering, such studies often compare their findings to the normally fluent speech of adults (Johnson, 1961). Thus, fluency measures of these Brazilian Portuguese adults can be compared to findings from other research and to evaluate their suitability as benchmarks for normal Brazilian Portuguese speakers (Ardila et al., 1994; Block & Killen, 1996). -

Even though this study included only 30 subjects, the obtained means may provide quantitative indexes that can be applied to both clinical and scientific projects if they are relatively consistent with findings from other studies.

Typical disfluent events predominated in the disfluencies found in the speech of this sample of adults, which agrees with the observations of Campell & Hill (1994) and Johnson (1961). In contrast, disfluencies considered to be less typical occur much less frequently as has been pointed out by Andrade (2000), Andrews et al. (1983), Van Riper and Emerick (1990), Wingate (1964), among others. However, the percentage of s :uttered sylla ales observed among this study’s subjects (0.2 - 0.7%), was less than half of the upper limit of normality (i.e., 2%) often reported in the literature (Bloodstein, 1979; Franken et al., 1995; Howell et al., 1997; Ingham & Riley, 1998; Kelly, 1994; Perellé, 1995; Ryan, 1992; Schwartz & Conture, 1988; Throneburg et al., 1994; Yairi & Ambrose, 1992; Yairi et al., 1993). It should be remembered that these studies involved speakers of a different language.

Comparisons of speech rate measures found in this study with those reported elsewhere reveal other discrepancies. For example, Lutz and Mallard (1986) found means of 212.4 SPM and 159.5 WPM for native English speakers in the U.S. In the present study, however, the means were 237.67 SPM and 128.79 WPM. This finding may be explained by differences in the methodologies used to analyze speech samples, as well as by differences between the two languages. For example, the grammatical word segmentation in English is completely different from that of Portuguese.

The measures obtained in this study can serve as guidelines during diagnostic procedures, in deciding when to dismiss a patient or in selecting subjects for control groups in research. Of course these measures should be used clinically along with other factors, such as a patient’s clinical background, presence of family history, self-perception of the problem, speech naturalness, among others. In this context, the speech of normal speaking adults in Brazilian Portuguese, in 200 syllable samples, is likely to be characterized by:

  1. 12 to 21 typical disfluent events;
  2. 0 to 2 less typical disfluent events;
  3. 7 to 10% speech discontinuity;
  4. 0.2 to 0.7% stuttered syllables; 380 Theory, research and therapy in fluency disorders
  5. 219 to 257 syllables per minute; f. 117 to 140 words per minute.
Although this study analyzed the speech samples of a reduced number of subjects, the effort to determine fluency parameters can be considered the first step towards understanding fluent speech of Brazilian Portuguese speaking adults. Therefore more studies in the area are necessary to confirm these findings.


Andrade, C.R.F. de (1999). Diagnostico e intervencdo precoce no tratamento das gagueira infantis. Carapicufba, Pr6-fono.

Andrade, C.R.F.de (2000). Processamento da fala: aspectos da fluéncia. Pro’-fono Revista de Atualizaccio Cientifica, 12(1), 69-71.

Andrews, G., Craig, A., Feyer, A.M., Hoddinot, S., Howie, P.& Neilson, M. (1983). Stuttering: a review of research findings and theories circa 1982. Journal of Speech and Hearing Disorder; 48, 226-246.

Ardila, A ., Bateman, J. R., Nifio, C. R., Pulido, E., Rivera, D. B. & Vanegas, C. J. (1994). An epidemiologic study of stuttering. Journal of Communication Disorders, 27, 37-48.

Barbosa, L.M.G.& Chiari, B.M. (1998). Gagueira: etiologia, prevenccio e tratamento. Carapicufba, Pro-fono. .

Block, S. & Killen, D. (1996). Speech rates of Australian English-speaking children and adults. Australian Journal of Human Communication Disorders, 24, 39-44.

Bloodstein, O. (1979): Speech pathology: an introduction, Boston, Houghton Mifflin Company, 103-159.

Campbell, J .& Hill, D. (1994). Systematic disfluency analysis. In: Stuttering therapy, (pp. 51-75). Memphis: Northwestern University & Stuttering Foundation of America.

Franken M.C., Boves, L., Peters,  & Webster, R.L. (1995). Perceptual rating instrument for speech evaluation of stuttering treatment. Journal of Speech and Hearing Research, 38, 280- 288.

Goldman-Eisler, F. (1961). The continuity of speech utterance: Its determinants and its significance. Language and Speech, 4, 220-231.

Howell, P., Sackin, S.& Glenn, K. (1997). Development of a two stage procedure for the automatic recognition of disfiuencies in the speech of children who stutter: I. Psychometric procedures appropriate for selection of training material for lexical dysfluency classifiers. Journal of Speech, Language and Hearing Research,_40, 1073-1084.

Ingham, J .C.& Riley, G. (1998). Guidelines for documentation of treatment efficacy for young children who stutter. Journal of Speech, Language and Hearing Research, 41, 753-770. Johnson, W. (1961). Measurements of oral reading and speaking rate and disfluency of adult male and female stutterers and nonstutterers. Journal of Speech and Hearing Disorders Monograph Supplement 7, 1-20.

Kelly, E.M. (1994). Speech rates and turn-taking behaviors of children who stutter and their fathers. Journal of Speech and Hearing Research, 37, 1284-1294.

Lutz, K.C.& Mallard, A.R. (1986). Disfluencies and rate of speech in young adult nonstutters. Journal of Fluency Disorders, 11, 307-316.

Pais, C.T. (1986). Elemento de fonologia estrutural. In: Manual de Linguistica. Sao Paulo, ed. Global (23 edigao), 9-80.

Perellé, J . (1995). Transtornos da fala. 5.ed. Trad. Elisabete Pereira de Souza, Marly Bezerra Canongia, Regina Elly Alves de Farias, Rosa Maria Ramos, Sueli Tomazini Cassal. 202p.

Perkins, W. H. (1983). The problem of definition: commentary on “stuttering Journal of Speech and Hearing Disorders, 48, 246-249. 

Rosner, B. (1986). Fundamentals of biostatistics. (2nd. ed.) Massachusetts, PWS Publishers. Ryan, B.P. (1992). Articulation, language, rate, and fluency characteristics of stuttering and nonstuttering preschool children. Journal of Speech and Hearing Research, 35, 333-342. Sander, E. (1961). Reliability of the Iowa speech disfluency Test. Journal of Speech and Hearing Disorders Monograph Supplement 7, 21-30.

Schwartz, H.D.& Conture, E.G. (1988). Subgrouping young stutterers: preliminary behavioral observations. Journal of Speech and Hearing Research, 31, 62-71.

Starkweather,C.W. & Givens-Ackerman, J. (1997). Stuttering. Austin,TX:Pro-Ed.

Throneburg, R.N., Yairi, E.& Paden, E.P. (1994). Relation between phonologic difficulty and the occurence of disfluencies in the early stage of stuttering. Journal of Speech and Hearing Research, 37, 504-509.

Van Riper, C.& Emerick, L. (1990). Speech correction: an introduction to speech pathology and audiology (8 ed). Englewood Cliffs, NJ : Prentice Hall.

Wingate, M.E. (1964). A standard definition of stuttering. Journal of Speech and Hearing Research, 29, 484-488.

Yairi, E.& Ambrose, N. (1992). A longitudinal study of stuttering in children: a preliminary report. Journal of Speech and Hearing Research, 35, 755-760.

Yairi, E., Ambrose, N.G.& Niermann, R. (1993). The early months of stuttering: a developmental study. Journal of Speech and Hearing Research, 36, 521-528.

Yairi, E. (1997). Disfluency characteristics of childhood stuttering. In R.F. Curlee & G.M. Siegel (Eds.): Nature and treatment of stuttering: New directions (pp. 49-78). Boston: Allyn & Bacon,.

Zackievicz, D.V. (l999).Avaliacc'io quantitativa das disfluéncias em indivfduos gagos e fiuentes (Master Dissertation, University of Sao Paulo, 1999). Area de Fisiopatologia Experimental. Faculdade de Medicina da Universidade de Sao Paulo.


This research was supported by a grant from - 97/03651-8 FAPESP (Fundaciio de Amparo £1 Pesquisa do Estado de Séio Paulo). 


In preparation for the 2018 World Congress the IFA is implementing Japanese translations of some pages on the site. Choosing Japanese below to see these translations.

Not all pages are translated, but you can use Google translate to see a machine translation using the switch below

Google Translate

Follow the Joint World Congress