Filled Pause
Research Center

Filled Pause
Research Center

Filled Pause
Research Center

Investigating 'um' and 'uh' and other hesitation phenomena

Investigating 'um' and 'uh' and other hesitation phenomena

Investigating 'um' and 'uh' and other hesitation phenomena

The founding of hesitation phenomenology

It's difficult to pinpoint exactly what is the earliest study on filled pauses because in the earliest days, there was no agreed-upon name for the phenomenon and some early researchers who referred to "pauses" in their work might very well have included what we now normally refer to as filled pauses in that nomination. Thus, the earliest days of such research is best known in the context of the slightly broader attention given to hesitation phenomena.

Howard Maclay and Charles Osgood's 1959 paper, “Hesitation Phenomena in Spontaneous English Speech”[1], is a highly accessible introduction to the topic and it's interesting that many papers today have literature reviews that largely parallel their paper in its thematic structure. Frieda Goldman-Eisler's “A Comparative Study of Two Hesitation Phenomena”[2] continues this with a look at what is now clearly distinguished as silent and filled pauses. Although Maclay and Osgood's paper was published earlier, Goldman-Eisler is often referred to the godparent of pausology, as evidenced by Hans Dechert and Manfred Raupach's 1980 tome, "Temporal Variables in Speech: Studies in Honour of Frieda Goldman-Eisler"[3]. George Mahl's work in the 1960's on the relation between filled pauses and anxiety, as in “Relationship of disturbances and hesitations in spontaneous speech to anxiety”[4] with Stanislav Kasl, actually suggests he might also have a claim to founder status.

Other work throughout the 1960s and into the early 1970s shows that the field was gradually establishing itself as various aspects of hesitation phenomena -- and specifically filled pauses - came into focus. Donald Boomer and Allen Dittmann's “Hesitation Pauses and Juncture Pauses in Speech”[5] carried various observations about the phenomena that set the stage for further study. Percy Tannenbaum and colleagues' “Word predictability in the environments of hesitations”[6] was an early corpus study done in the days before corpora were big and digitized. They did it the old-fashioned way: by hand. James Martin and Winifred Strange looked at the perceptual side of hesitation phenomena in “The perception of hesitation in spontaneous speech”[7].

As 1970 approaches and passes, we see that the study of filled pauses is becoming more defined and methods more sophisticated. Mansur Lallgee and Mark Cook's “An experimental investigation of the function of filled pauses in speech”[8] is exactly what its title describes and perhaps one of the first filled pause studies so clearly described. Mark Cook then looks at filled pauses in relation to syntactic structure in “The incidence of filled pauses in relation to part of speech” as does Frieda Goldman-Eisler in “Pauses, Clauses, Sentences”[9].

By this time, the work in the field was becoming well-enough established that it was time for a summary work. This was provided by Sherry Rochester in "The significance of pauses in spontaneous speech"[10], an excellent overview of the preceding two decades' worth of work on the topic.

Much of the work in those early days has now been surpassed by newer work, but it's quite interesting to read the work in those days and see how the early researchers struggled with an interesting phenomenon, but with little data and lots of conjecture about it.


  • Howard Maclay, and Charles Osgood, “Hesitation Phenomena in Spontaneous English Speech,” Word, vol. 15, 1959, pp. 19-44. DOI: 10.1080/00437956.1959.11659682.

    Abstract This paper reports an exploratory investigation of hesitation phenomena in spontaneously spoken English. Following a brief review of the literature bearing on such phenomena, a quantitative study of filled and unfilled pauses, repeats, and false starts in the speech of some twelve participants in a conference is described. Analysis in terms of both individual differences and linguistic distribution is made, and some psycholinguistic implications are drawn, particularly as to the nature of encoding units and their relative uncertainty. A distinction between non-chance statistical dependencies and all-or-nothing dependencies in linguistic methodology is made.

  • Frieda Goldman-Eisler, “A Comparative Study of Two Hesitation Phenomena,” Language and Speech, vol. 4, no. 1, 1961, pp. 18-26. DOI: 10.1177/002383096100400102.

    Abstract The durations of hesitation devices such as the sounds /α, ∊, æ, r, ∂, m/, also called filled pauses, were measured and compared with the durations of silent hesitations or unfilled pauses. Their individual consistency and psychological significance were also investigated and the relation to uncertainty of filled pauses and unfilled pauses respectively was compared. It appears that under certain conditions of speech production the two hesitation phenomena reflect different internal processes.

  • Temporal Variables in Speech: Studies in Honour of Frieda Goldman-Eisler. The Hague: Mouton.1980. DOI:

    Abstract (none)

  • Stanislav V Kasl, and George F Mahl, “Relationship of disturbances and hesitations in spontaneous speech to anxiety,” Journal of Personality and Social Psychology, vol. 1, 1965. DOI: 10.1037/h0021918.

    Abstract Past work has indicated that flustered or confused speech can be classed into several distinct speech disturbance categories. Such disturbances, occurring frequently in everyday conversation, have no conventional semantic function. In the present study, 25 experimental and 20 control male Ss were used. Anxiety was manipulated in an interview setting. Under anxiety, the frequency of all speech disturbances, except the familiar "ah," showed a sizable increase. The frequency of ah’s increased strikingly in a change from normal to a telephonelike conversation. Such change did not affect the other disturbances. Measurement of palmar sweat revealed modest positive association with the speech disturbances. Exploration of the relationship of the Taylor MA scale to the disturbances suggested that the ah is functionally distinct from the other speech disturbances. (PsycINFO Database Record (c) 2016 APA, all rights reserved)

  • Donald Boomer, and Allen Dittmann, “Hesitation Pauses and Juncture Pauses in Speech,” Language and Speech, vol. 5, 1962, pp. 215-220. DOI: 10.1177/002383096200500404.

    Abstract A psychophysical comparison of speech pause perception thresholds for juncture and hesitation pauses yielded significantly lower thresholds for the latter. On the basis of these and other data, a functional and methodological distinction between these two types of pause is proposed.

  • Percy H. Tannenbaum, Frederick Williams, and Carolyn S. Hillier, “Word predictability in the environments of hesitations,” Journal of Verbal Learning and Verbal Behavior, vol. 4, no. 2, 1965, pp. 134 - 140. DOI: 10.1016/S0022-5371(65)80097-4.

    Abstract Two experiments were conducted to study the predictability of words in hesitation contexts. The first study focused on a comparison of the first word after hesitations with words sampled from fluent contexts. The second study involved gathering predictability data for all words in a language sample. Results supported the hypothesis that words subsequent to hesitations tend to be less predictable than words uttered in fluent context. But the associated hypothesis that the word antecedent to hesitations is more predictable than other fluent context was not supported. This led to further analysis of predictability of words in the environments of different hesitations, specifically filled pauses and repeats. The implication drawn was that different types of hesitations index different kinds of encoding decision points.

  • James Martin, and Winifred Strange, “The perception of hesitation in spontaneous speech,” Psychonomic Journals: Perception & Psychophysics, vol. 3, no. 6, November 1968, pp. 427-438. DOI: 10.3758/BF03205750.

    Abstract The issue in this paper was whether attending to acoustic elements and to message elements in a speech signal were compatible operations. In four experiments Ss listened for pauses and other hesitation phenomena in spontaneous speech; in three the task was reproduction of heard speech to include hesitations; in one the task was simply the marking of heard hesitations on transcripts. Experimental variables were instructions, degree of “ungrammaticality” of hesitations in speech inputs, time interval between listening and reproduction, and task manipulations along a continuum between simple hesitation detection and hesitation detection plus simultaneous speech decoding. Results were: (I) In all experiments Ss displaced within-constituent hesitations to constituent boundaries, suggesting a grammatical organization between input and output. (2) Instructional set to reproduce hesitations increased hesitations and words but at the expense of per cent words correct, suggesting that attending to acoustic elements such as hesitations was an interfering task during speech decoding. (3) The hesitation shift persisted in the hesitation-marking task when simultaneous speech decoding was required by the nature of the task, indicating that speaking (encoding) characteristics may not completely account for the shift. (4) The distribution of hesitation marking errors toward grammatical organization seemed to require an account in terms of perceptual processes during listening.

  • Mansur Lallgee, and Mark Cook, “An experimental investigation of the function of filled pauses in speech,” Language and Speech, vol. 12, no. 1, January 1969, pp. 24-29. DOI: 10.1177/002383096901200102.

    Abstract Filled pauses have been described as a product of anxiety, and have also been explained as attempts by the speaker to maintain control of the ’ floor’. The latter hypothesis is tested directly, by altering the pressure on the subject to continue speaking. Possible confounding effects of anxiety are controlled for. Filled pauses do not increase, as pressure to continue speaking increases. It is suggested that the ’ control’ hypothesis may apply only to monologues ; evidence concerning the relative frequency of filled pauses in monologues and dialogues is presented.

  • Frieda Goldman-Eisler, “Pauses, Clauses, Sentences,” Language and Speech, vol. 15, no. 2, April 1972, pp. 103-113. DOI: 10.1177/002383097201500201.

    Abstract The tool of pause measurement was applied to the question of the psychological reality of syntactic structures in spontaneous speech. The material investigated covered a wide field of speech productions, of different speakers and different speech tasks. Their analysis showed that the hierarchy of syntactic structures is reflected differentially in the pause structure of spontaneous speech. When readings of the spontaneous texts were compared with the original spontaneous speech it emerged that the reading process modifies the pausing for different syntactic structures differently. Sentences as distinct from clauses are marked by their temporal cohesion in spontaneous speech as well as in reading. This fact is discussed with reference to Wundt’s analytical theory of sentence-wholes.

  • Sherry R. Rochester, “The significance of pauses in spontaneous speech,” Journal of Psycholinguistic Research, vol. 2, no. 1, 1973, pp. 51-81. DOI: 10.1007/BF01067111.

    Abstract Studies of filled and silent pauses performed in the last two decades are reviewed in order to determine the significance of pauses for the speaker. Following a brief history, the theoretical implications of pause location are examined and the relevant studies summarized. In addition, the functional significance of pauses is considered in terms of cognitive, affective-state, and social interaction variables.